Prey-predator dynamics with additional food and Holling type-III response: application to Bemisia tabaci Authors: D. Bhanu Prakash, Ch Chaitanya, D. K. K. Vamsi, Ahmad Aloqaily, Nabil Mlaiki Published: 2026 ABSTRACT Interactions between pests and their natural enemies play a central role in ecological balance and sustainable pest management. In the present work, we examine a prey–predator system in which the predator population experiences intra-specific competition and receives an external food supplement. The feeding mechanism of the predator follows a Holling type-III functional response. We demonstrate that the proposed formulation admits a unique, globally positive solution and investigate the stability of its equilibria. Analytical exploration reveals the occurrence of distinct bifurcation phenomena, including transcritical and saddle-node transitions. Numerical investigations further illustrate the emergence of hysteresis loops, indicating bistability within specific parameter ranges. Experimental observations concerning the predation of Bemisia tabaci by Orius albidipennis validate the Holling type-III assumption and serve to parameterize the theoretical model. The findings underscore the dual impact of additional food and predator competition in shaping biological control outcomes and improving pest suppression efficiency. Introduction Biological control, which employs living organisms to regulate pest populations, is a vital component of environmentally responsible pest management. Traditional chemical pesticides often yield quick results but introduce ecological and health-related side effects. In contrast, biocontrol methods aim to maintain agricultural productivity while preserving ecological integrity1. However, the success of such methods depends on the complex dynamics between pests, their predators, and the surrounding environment. Without careful design, even a beneficial intervention can destabilize an ecosystem and cause the outbreak or extinction of species. Mathematical modelling provides a quantitative approach to understanding these interactions. Prey–predator systems capture how populations evolve through mutual influences, enabling researchers to identify stable coexistence regimes, thresholds for extinction, and possible oscillatory behaviours. Among various forms of predator feeding responses, the Holling type-III (sigmoidal) form has received particular attention. It characterizes situations where predation is weak at low prey densities due to predator learning or prey refuge effects and accelerates as prey become more abundant. Such behaviour is commonly observed in insect predators, sometimes referred to as “lazy predators”, which respond more actively only when prey are sufficiently available. A growing body of research has refined these models by incorporating ecological mechanisms such as prey grouping, predator interference, harvesting pressure, and the influence of additional food resources2–5. The inclusion of non-reproductive supplemental food alters predator survival and functional response characteristics, which can either stabilize or destabilize the ecosystem depending on food quality and abundance6,7. Analytical and numerical studies in this area have revealed that these factors often lead to bifurcations, multi-stability, or even chaotic transitions, emphasizing the need for rigorous mathematical treatment8–10. Beyond qualitative dynamics, several works have explored control-oriented formulations of predator–prey interactions. Optimal and time-optimal control approaches aim to determine effective intervention strategies, such as adjusting food supplementation or predator release rates, to reach desired population levels while minimizing time or cost11,12. These frameworks bridge theoretical ecology and practical pest management by converting ecological processes into systems amenable to optimization. To enhance sustainability in pest suppression, research increasingly focuses on ecological modeling and predator-based control mechanisms. The management of key pests such as Tetranychus urticae using biological agents under controlled conditions demonstrates the potential for integrating multiple natural enemies within polyhouse systems13. Mathematical approaches have further advanced understanding by incorporating factors such as additional food resources and interspecific competition into prey-predator models, yielding refined predictions of system stability and control efficacy14. Studies on eco-epidemiological stability15, predator harvesting effects in stochastic environments16, and delay-induced dynamics in Leslie-Gower models17 contribute to the theoretical foundation of these systems. Furthermore, bifurcation analyses of additional food-provided systems with anti-predator behavior extend this understanding, offering valuable insights into the nonlinear mechanisms underlying biological control efficiency18. From an applied perspective, Bemisia tabaci (whitefly) is a significant agricultural pest responsible for severe yield losses in multiple crops. Its biological suppression through predatory bugs such as Orius albidipennis has been studied experimentally, with the predator exhibiting a type-III functional response to prey density19. This feature makes the system ideal for examining the combined roles of predator competition and supplemental feeding in maintaining long-term stability. The combination of these three mechanisms—intra-specific competition among predators, provision of supplementary food, and a Holling type-III (sigmoidal) functional response—is ecologically well-motivated. First, predatory arthropods such as Orius spp. are known to exhibit interference and contest competition at high densities, which limits their per-capita consumption rate10. Second, supplementary food (e.g., pollen, factitious prey, or sugar solutions) is a standard augmentative strategy that sustains predator populations during periods of low pest density7. Third, the sigmoidal (type-III) response captures prey-switching and learning behavior frequently observed in generalist predators: at low pest densities predation is negligible, but it accelerates sharply once a density threshold is crossed. Individually, each mechanism has been studied; however, their combined effect—particularly how supplementary food interacts with intra-specific competition under a sigmoidal response—remains unexplored. This gap motivates the present formulation. Recent studies further support the relevance of these combined mechanisms. Mishra and Parshad20 demonstrated that cannibalistic interactions among natural enemies, when coupled with additional food, can produce qualitatively different dynamics than either mechanism alone. Similarly, Mishra et al.21 showed that harvesting interventions in additional-food systems lead to complex bifurcation structures including imperfect bifurcations22. These findings underscore the importance of studying multiple ecological mechanisms simultaneously. The present study formulates and analyses a prey–predator system that includes intra-specific competition among predators and external food support. The main objectives are to identify equilibrium states, assess their stability, and detect possible bifurcation scenarios. The investigation further incorporates a time-optimal control problem where the quality and quantity of the additional food act as controllable variables. Such a framework provides insights into how auxiliary feeding strategies can enhance biological control efficiency without disrupting ecosystem balance. The principal contributions of this work are as follows: We formulate and rigorously analyse, for the first time, a prey–predator model that simultaneously incorporates a Holling type-III functional response, intra-specific competition among predators, and a non-reproductive additional food source. While each of these mechanisms has been studied individually2–5, their combined effect—particularly the interplay between supplementary feeding and predator competition under a sigmoidal response—has not been previously investigated.We provide a complete analytical framework including: (a) proof of global boundedness and positivity, (b) explicit existence conditions for all equilibria including the interior coexistence state governed by a fifth-degree polynomial, (c) stability analysis with ecological interpretation of all threshold conditions, and (d) rigorous proofs of transcritical and saddle-node bifurcations via Sotomayor’s theorem.We demonstrate the emergence of hysteresis loops with respect to intra-specific competition, revealing that the system’s long-term state depends on the history of parameter variation—a finding with direct implications for biocontrol strategy design.We validate the type-III functional response assumption using experimental data from the Orius albidipennis–Bemisia tabaci system19, bridging theoretical predictions and empirical observations.The remainder of this article is organized as follows. Section 2 presents the formulation of the proposed mathematical model and introduces the parameters of interest. Section 3 and Section 4 addresses the existence and stability of equilibria. Section 5 and section 6 examines the occurrence of local and global bifurcations, supported by numerical simulations. Section 7 explores the ecological implications and optimal control results, while Section 8 discusses the applications of this research to a pest Bemisia tabaci. Section 9 summarizes key conclusions and potential directions for future work. We formulate and rigorously analyse, for the first time, a prey–predator model that simultaneously incorporates a Holling type-III functional response, intra-specific competition among predators, and a non-reproductive additional food source. While each of these mechanisms has been studied individually2–5, their combined effect—particularly the interplay between supplementary feeding and predator competition under a sigmoidal response—has not been previously investigated. We formulate and rigorously analyse, for the first time, a prey–predator model that simultaneously incorporates a Holling type-III functional response, intra-specific competition among predators, and a non-reproductive additional food source. While each of these mechanisms has been studied individually2–5, their combined effect—particularly the interplay between supplementary feeding and predator competition under a sigmoidal response—has not been previously investigated. We provide a complete analytical framework including: (a) proof of global boundedness and positivity, (b) explicit existence conditions for all equilibria including the interior coexistence state governed by a fifth-degree polynomial, (c) stability analysis with ecological interpretation of all threshold conditions, and (d) rigorous proofs of transcritical and saddle-node bifurcations via Sotomayor’s theorem. We provide a complete analytical framework including: (a) proof of global boundedness and positivity, (b) explicit existence conditions for all equilibria including the interior coexistence state governed by a fifth-degree polynomial, (c) stability analysis with ecological interpretation of all threshold conditions, and (d) rigorous proofs of transcritical and saddle-node bifurcations via Sotomayor’s theorem. We demonstrate the emergence of hysteresis loops with respect to intra-specific competition, revealing that the system’s long-term state depends on the history of parameter variation—a finding with direct implications for biocontrol strategy design. We demonstrate the emergence of hysteresis loops with respect to intra-specific competition, revealing that the system’s long-term state depends on the history of parameter variation—a finding with direct implications for biocontrol strategy design. We validate the type-III functional response assumption using experimental data from the Orius albidipennis–Bemisia tabaci system19, bridging theoretical predictions and empirical observations. We validate the type-III functional response assumption using experimental data from the Orius albidipennis–Bemisia tabaci system19, bridging theoretical predictions and empirical observations. Model formulation The prey–predator framework considered here accounts for two main ecological effects: intra-specific competition among predators and the provision of an additional, non-reproductive food source. The predator’s feeding rate follows the Holling type-III response, which captures the saturation of predation at high prey densities and a low consumption rate when prey are scarce. The proposed prey-predator system is given by1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \begin{aligned} \frac{\textrm{d} N}{\textrm{d} T}&= r N \left( 1-\frac{N}{K} \right) -\frac{c N^2 P}{a^2+N^2+\alpha \eta A^2}, \\ \frac{\textrm{d} P}{\textrm{d} T}&= \delta _1 \left( \frac{N^2 + \eta A^2}{a^2+N^2+\alpha \eta A^2} \right) P - m_1 P - d P^2. \end{aligned} \end{aligned}$$\end{document}Let the prey population grow logistically in the absence of predation, while predators consume prey and also benefit from supplementary food. Non-dimensionalization We introduce the following substitutions to reduce the number of free parameters. Let\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} N = a\,x, \quad P = \frac{ar}{c}\,y, \quad t = r\,T, \end{aligned}$$\end{document}where a is the half-saturation constant, r the prey intrinsic growth rate, and c the maximum predation rate. Then\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \frac{dN}{dT} = r\,a\,\frac{dx}{dt}, \qquad \frac{dP}{dT} = \frac{ar}{c}\,r\,\frac{dy}{dt}. \end{aligned}$$\end{document}Substituting into the prey equation:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} r\,a\,\frac{dx}{dt}&= r\,(ax)\!\left( 1 - \frac{ax}{K}\right) - \frac{c\,(ax)^{2}\,\dfrac{ar}{c}\,y}{a^{2} + (ax)^{2} + \alpha \,\eta \,A^{2}} \\&= r\,a\,x\!\left( 1 - \frac{x}{\gamma }\right) - \frac{r\,a\,x^{2}\,y}{1 + x^{2} + \alpha \,\xi }, \end{aligned}$$\end{document}where γ = K/a and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi = \eta \,(A/a)^{2}$$\end{document}. Dividing both sides by r a gives\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \frac{dx}{dt} = x\!\left( 1 - \frac{x}{\gamma }\right) - \frac{x^{2}\,y}{1 + x^{2} + \alpha \,\xi }. \end{aligned}$$\end{document}Similarly, for the predator equation:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \frac{ar}{c}\,r\,\frac{dy}{dt}&= \delta _{1}\!\left( \frac{(ax)^{2} + \eta \,A^{2}}{a^{2} + (ax)^{2} + \alpha \,\eta \,A^{2}}\right) \!\frac{ar}{c}\,y - m_{1}\,\frac{ar}{c}\,y - d\!\left( \frac{ar}{c}\,y\right) ^{\!2}. \end{aligned}$$\end{document}Dividing by \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\dfrac{ar^{2}}{c}$$\end{document} and defining \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta = \delta _{1}/r$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m = m_{1}/r$$\end{document}, ε = da/c:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \frac{dy}{dt} = \frac{\delta \,(x^{2}+\xi )\,y}{1+x^{2}+\alpha \,\xi } - m\,y - \epsilon \,y^{2}. \end{aligned}$$\end{document}The dimensionless parameters and their ecological meanings are summarised in Table 1.Table 1Dimensionless parameters and their ecological interpretation.ParameterDefinitionEcological meaningγK/aScaled carrying capacityξ\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\eta \,(A/a)^{2}$$\end{document}Scaled quantity of additional foodα–Quality of additional foodδ\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta _{1}/r$$\end{document}Scaled conversion efficiencym\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m_{1}/r$$\end{document}Scaled predator mortalityεda/cScaled intra-specific competition Table 1Dimensionless parameters and their ecological interpretation.ParameterDefinitionEcological meaningγK/aScaled carrying capacityξ\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\eta \,(A/a)^{2}$$\end{document}Scaled quantity of additional foodα–Quality of additional foodδ\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta _{1}/r$$\end{document}Scaled conversion efficiencym\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m_{1}/r$$\end{document}Scaled predator mortalityεda/cScaled intra-specific competition Dimensionless parameters and their ecological interpretation. After introducing appropriate non-dimensional variables \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$N = a x,\ P=\frac{ar y}{c},\ t = r T$$\end{document}, the governing equations are represented in a reduced form as follows.2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \begin{aligned} \frac{\textrm{d} x}{\textrm{d} t}&= x \left( 1 - \frac{x}{\gamma } \right) - \frac{x^2 y}{1 + x^2 + \alpha \xi }, \\ \frac{\textrm{d} y}{\textrm{d} t}&= \frac{\delta \left( x^2 + \xi \right) y}{1+x^2+\alpha \xi } - m y - \epsilon y^2, \end{aligned} \end{aligned}$$\end{document}where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\gamma = \frac{K}{a},\ \xi = \eta (\frac{A}{a})^2,\ \epsilon = \frac{d a}{c},\ m = \frac{m_1}{r},\ \delta = \frac{\delta _1}{r}$$\end{document}. Each parameter possesses an ecological interpretation: the parameters \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$r,K,\delta ,m,\alpha ,\xi$$\end{document} represents the prey’s intrinsic growth rate, its carrying capacity, the predator’s conversion efficiency, mortality rate, and the qualitative and quantitative measures of the additional food respectively. The dimensionless model thus reflects how both biological and artificial resources influence predator growth and prey suppression. To simplify the analysis, the system is expressed in a normalized time scale. This transformation allows us to investigate its qualitative behaviour without dependence on particular units, facilitating bifurcation and stability analyses presented later. Before proceeding to the analysis of equilibria and their stability, we first establish the positivity and boundedness of solutions to ensure that the model remains biologically meaningful and the unrealistic population blow-ups are excluded. Theorem 2.1 The system (2) admits bounded, positive solutions for all admissible initial conditions \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(x(0), y(0)) \in {\mathbb {R}}^{2}_{+}$$\end{document}. Proof Positivity. We show that the positive quadrant \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathbb {R}}^{2}_{+}$$\end{document} is positively invariant. On the boundary x = 0:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \left. \frac{dx}{dt}\right| _{x=0} = 0, \end{aligned}$$\end{document}so x = 0 is an invariant set; trajectories starting with x(0)> 0 cannot reach x = 0 in finite time. On the boundary y = 0:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \left. \frac{dy}{dt}\right| _{y=0} = 0, \end{aligned}$$\end{document}so y = 0 is likewise invariant. Hence, solutions starting in the open positive quadrant remain there for all t> 0. Boundedness. Define the auxiliary function \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$W(t) = x(t) + \frac{1}{\delta }\,y(t)$$\end{document}. Then\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \frac{dW}{dt}&= x\!\left( 1 - \frac{x}{\gamma }\right) - \frac{x^{2}y}{1+x^{2}+\alpha \xi } + \frac{1}{\delta }\!\left[ \frac{\delta (x^{2}+\xi )y}{1+x^{2}+\alpha \xi } - my - \epsilon y^{2}\right] \\&= x\!\left( 1 - \frac{x}{\gamma }\right) - \frac{x^{2}y}{1+x^{2}+\alpha \xi } + \frac{(x^{2}+\xi )y}{1+x^{2}+\alpha \xi } - \frac{m}{\delta }y - \frac{\epsilon }{\delta }y^{2} \\&= x\!\left( 1 - \frac{x}{\gamma }\right) + \frac{\xi \,y}{1+x^{2}+\alpha \xi } - \frac{m}{\delta }y - \frac{\epsilon }{\delta }y^{2}. \end{aligned}$$\end{document}Since \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{\xi }{1+x^{2}+\alpha \xi } \le \frac{\xi }{1+\alpha \xi } \le \frac{1}{\alpha }$$\end{document} (for α> 0), we have\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \frac{dW}{dt} \le x\!\left( 1 - \frac{x}{\gamma }\right) + \frac{y}{\alpha } - \frac{m}{\delta }y - \frac{\epsilon }{\delta }y^{2}. \end{aligned}$$\end{document}For any σ> 0, adding σ W to both sides:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \frac{dW}{dt} + \sigma W \le x\!\left( 1+\sigma - \frac{x}{\gamma }\right) + \left( \frac{1}{\alpha } - \frac{m}{\delta } + \frac{\sigma }{\delta }\right) y - \frac{\epsilon }{\delta }y^{2}. \end{aligned}$$\end{document}Choosing \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\sigma < m - \delta /\alpha$$\end{document} (which requires \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m> \delta /\alpha$$\end{document}; when this fails, one can use a modified Lyapunov function W = x + cy for suitable c), the right-hand side is bounded above by a constant M> 0. By the comparison lemma,\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} W(t) \le \frac{M}{\sigma } + \left( W(0) - \frac{M}{\sigma }\right) e^{-\sigma t} \le \max \!\left\{ W(0),\, \frac{M}{\sigma }\right\} , \end{aligned}$$\end{document}so W(t), and hence x(t) and y(t), remain bounded for all t \ge 0. □ Existence of equilibria The prey and predator nullclines of the system (2) are given by \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$x = 0, \ \ 1- \frac{x}{\gamma } - \frac{x y}{1+x^2 + \alpha \xi } = 0$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$y = 0, \ \ \frac{\delta (x^2 + \xi )}{1+x^2+ \alpha \xi } - m - \epsilon y = 0$$\end{document} respectively.Fig. 1Nullcline configurations for system (2). Red curves: prey nullclines; blue curves: predator nullclines; green curves: prey nullcline in the absence of additional food. Panels A–D illustrate the existence of interior equilibria even by varying the qualitative behavior of the nullclines. Dashed vertical line indicates the carrying capacity γ. Nullcline configurations for system (2). Red curves: prey nullclines; blue curves: predator nullclines; green curves: prey nullcline in the absence of additional food. Panels A–D illustrate the existence of interior equilibria even by varying the qualitative behavior of the nullclines. Dashed vertical line indicates the carrying capacity γ. Graphical examination of the prey and predator nullclines are presented in Figure 1. In all the possible scenarios, the system presents possible intersections, suggesting scenarios with at least one coexisting equilibrium point depending on food quality, food quantity, and intra-specific competition. Analytical conditions ensuring the existence of positive equilibria can be derived from these relations.\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned}\delta \xi - m(1+\alpha \xi ) \le 0, \ \sqrt{\frac{\delta \xi - m(1+\alpha \xi )}{m - \delta }} < \gamma \implies 0> \delta \xi - m(1+\alpha \xi )> - \gamma ^2 (\delta - m).\end{aligned}$$\end{document}The equilibria correspond to points where prey and predator populations remain constant over time. These arise when the rates of change of both variables vanish. Besides the trivial extinction point \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(E_0 = (0,0))$$\end{document}, two boundary equilibria emerge–one representing the persistence of prey without predators \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(E_1 = (\gamma , 0))$$\end{document}, and another describing a predator population sustained solely by additional food \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\left( E_2 = \left( 0,\frac{\delta \xi - m (1+\alpha \xi )}{\epsilon (1+\alpha \xi )}\right) \right)$$\end{document}. Under biologically feasible parameter combinations, the following interior coexistence equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(E^* = (x^*, y^*))$$\end{document} also exists where both populations survive.3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \centering y^* = \frac{(\delta - m) (x^*)^2 + \delta \xi - m (1+\alpha \xi )}{\epsilon (1 + (x^*)^2 + \alpha \xi )}, \end{aligned}$$\end{document}4\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \begin{aligned} \frac{\epsilon }{\gamma } (x^*)^5 - \epsilon (x^*)^4 + \left( \delta - m + \frac{2 \epsilon }{\gamma } (1 + \alpha \xi ) \right) (x^*)^3&\\ - 2 \epsilon (1+\alpha \xi ) (x^*)^2 + \left( \frac{\epsilon }{\gamma } (1+\alpha \xi )^2 + \delta \xi - m (1+\alpha \xi ) \right) (x^*) - \epsilon (1+\alpha \xi )^2&= 0. \end{aligned} \end{aligned}$$\end{document} Feasibility of interior equilibria Equation (4) is a fifth-degree polynomial in \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$x^{*}$$\end{document}. By Descartes’ rule of signs, we can bound the number of positive real roots. Examining the sign pattern of the coefficients:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \frac{\epsilon }{\gamma }\,x^{5} - \epsilon \,x^{4} + \left( \delta - m + \frac{2\epsilon }{\gamma }(1+\alpha \xi )\right) x^{3} - 2\epsilon (1+\alpha \xi )\,x^{2} + \left( \frac{\epsilon }{\gamma }(1+\alpha \xi )^{2} + \delta \xi - m(1+\alpha \xi )\right) x - \epsilon (1+\alpha \xi )^{2}. \end{aligned}$$\end{document}Since \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon ,\gamma> 0$$\end{document}, the leading coefficient is positive and the constant term is negative. The signs of the intermediate coefficients of the powers of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$x^4$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$x^2$$\end{document} are negative. In typical ecological scenarios where δ> m (predator conversion exceeds mortality), the sign of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$x^3$$\end{document} is positive. Only the sign of x depends depend on parameter values. Therefore, the sign sequence is (+, -, +, -, ± , -), yielding at most five sign changes, and hence at most five positive real roots. However, biological admissibility further restricts the roots: a root \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$x^{*}$$\end{document} must satisfy both\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \frac{\epsilon }{1 + \epsilon /\gamma }< x^{*} < \gamma \qquad \text {and}\qquad y^{*} = \frac{(\delta -m)(x^{*})^{2} + \delta \xi - m(1+\alpha \xi )}{\epsilon (1 + (x^{*})^{2} + \alpha \xi )}> 0. \end{aligned}$$\end{document}The first condition ensures that the prey equilibrium lies between the minimum viable population and the carrying capacity; the second ensures a positive predator population. Numerical exploration across extensive parameter grids (see Figure 6) indicates that the system typically admits one, three, or five biologically admissible interior equilibria, with the bistable regime (three equilibria: two stable nodes/foci separated by a saddle) being of greatest ecological relevance. Putting all these results together, we have the following theorem. Lemma 3.1 The system (2) admits an interior equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E^* = (x^*,y^*)$$\end{document} which is a solution of the equations (3) and (4) that satisfies the conditions \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta \xi - m(1+\alpha \xi )> - \gamma ^2 (\delta - m)$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{\epsilon }{1 + \frac{\epsilon }{\gamma }}< x^* < \gamma$$\end{document}. Stability of equilibria To understand the long-term outcomes near each equilibrium, the system (2) is linearized and the corresponding Jacobian matrix is evaluated at equilibrium points. The eigenvalues of this matrix determine whether small perturbations decay or amplify.\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned}J = \begin{bmatrix} \frac{\partial }{\partial x} f(x,y) & \frac{\partial }{\partial y} f(x,y)\\ \frac{\partial }{\partial x} g(x,y) & \frac{\partial }{\partial y} g(x,y) \end{bmatrix},\end{aligned}$$\end{document}where\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} f(x,y)= & x \left( 1-\frac{x}{\gamma } \right) - \frac{x^2y}{1+x^2+\alpha \xi }, \\ g(x,y)= & \delta \left( \frac{x^2+\xi }{1 + x^2 + \alpha \xi } \right) y - m y - \epsilon y^2, \end{aligned}$$\end{document}and\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \frac{\partial }{\partial x} f(x,y)= & 1 - \frac{2 x}{\gamma } - \frac{2 x y (1 + \alpha \xi )}{(1 + x^2 + \alpha \xi )^2}, \\ \frac{\partial }{\partial y} f(x,y)= & -\frac{x^2}{1 + x^2 + \alpha \xi }, \\ \frac{\partial }{\partial x} g(x,y)= & \frac{2 \delta x y (1+(\alpha - 1) \xi )}{(1+x^2 + \alpha \xi )^2}, \\ \frac{\partial }{\partial y} g(x,y)= & \frac{\delta (x^2 + \xi )}{1+x^2+\alpha \xi } - m - 2 \epsilon y. \end{aligned}$$\end{document}At the trivial equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_0 = (0,0)$$\end{document}, we obtain the jacobian as\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} J\left( E_0 \right) = \begin{bmatrix} 1 & 0 \\ 0 & \frac{\delta \xi - m (1+ \alpha \xi )}{1+\alpha \xi } \end{bmatrix}. \end{aligned}$$\end{document}The eigenvalues of this jacobian matrix are \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$1, \frac{\delta \xi - m (1+ \alpha \xi )}{1+\alpha \xi }$$\end{document}. If \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta \xi - m (1+ \alpha \xi )> 0$$\end{document}, then both the eigenvalues have same signs. This makes the equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_0 = (0,0)$$\end{document} unstable. If \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta \xi - m (1+ \alpha \xi ) < 0$$\end{document}, then both the eigenvalues will have opposite signs. This makes the point \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_0 = (0,0)$$\end{document} a saddle point. In the absence of additional food, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_0 = (0,0)$$\end{document} is a saddle point. At the axial equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1 = (\gamma ,0)$$\end{document}, we obtain the jacobian as\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} J(E_1) = \begin{bmatrix} -1 & \frac{-\gamma ^2}{1+ \gamma ^2 + \alpha \xi } \\ 0 & \frac{(\delta - m) \gamma ^2 + \delta \xi - m (1+\alpha \xi )}{\gamma ^2 + 1 + \alpha \xi } \end{bmatrix}. \end{aligned}$$\end{document}The eigenvalues of this jacobian matrix are \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$-1, \frac{(\delta - m) \gamma ^2 + \delta \xi - m (1+\alpha \xi )}{\gamma ^2 + 1 + \alpha \xi }$$\end{document}. If \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(\delta - m) \gamma ^2 + \delta \xi - m (1+\alpha \xi )> 0$$\end{document}, then both the eigenvalues will have opposite sign resulting in a saddle point. If \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(\delta - m) \gamma ^2 + \delta \xi - m (1+\alpha \xi ) < 0$$\end{document}, then it will be an asymptotically stable node. In the absence of additional food, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1 = (\gamma , 0)$$\end{document} is a saddle point if \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(\delta - m) \gamma ^2> m$$\end{document}. Else it is a stable equilibrium. We now consider another axial equilibrium which exists only for the additional food provided system (2). At this axial equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_2 = \left( 0, \frac{\delta \xi - m (1+\alpha \xi )}{\epsilon (1 + \alpha \xi )}\right)$$\end{document}, the associated jacobian matrix is given as\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} J(E_2) = \begin{bmatrix} 1 & 0 \\ 0 & -\frac{\delta \xi - m (1+\alpha \xi )}{1 + \alpha \xi } \end{bmatrix}. \end{aligned}$$\end{document}Since this equilibrium exists in positive xy-quadrant only when \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta \xi - m (1+\alpha \xi )> 0$$\end{document}, the eigenvalues of the associated jacobian matrix are of opposite side which resulting in a saddle equilibrium. The following lemmas present the stability nature of the trivial and axial equilibria. Lemma 4.1 The trivial equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_0 = (0,0)$$\end{document} is saddle (unstable node) if\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \delta \xi - m (1+\alpha \xi ) < (>)\ 0. \end{aligned}$$\end{document} Remark 1 (Ecological interpretation) The condition \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta \xi - m(1+\alpha \xi ) < 0$$\end{document} can be rewritten as \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{\delta \xi }{1+\alpha \xi } < m$$\end{document}, which means that the effective energy gain from additional food alone (left-hand side) is insufficient to offset predator mortality m. When this holds, predators cannot survive on additional food alone, and the extinction state is a saddle—prey can escape the neighbourhood of (0, 0) while predators decline. Conversely, if \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{\delta \xi }{1+\alpha \xi }> m$$\end{document}, the additional food is rich enough to sustain predators even without prey, making both populations grow away from extinction (unstable node). Lemma 4.2 The axial equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1 = (\gamma ,0)$$\end{document} is stable node (saddle) if\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \delta \xi - m (1+\alpha \xi ) < (>)\ - (\delta - m) \gamma ^2. \end{aligned}$$\end{document} Remark 2 (Ecological interpretation) The threshold \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta \xi - m(1+\alpha \xi ) = -(\delta -m)\gamma ^{2}$$\end{document} determines whether predators can invade a prey-only community at carrying capacity. When \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{\delta (\gamma ^{2}+\xi )}{1+\gamma ^{2}+\alpha \xi } < m$$\end{document} (the prey-only state is stable), the combined prey availability at x=γ and supplementary food do not provide sufficient energy for predator establishment. Biologically, this corresponds to situations where the predator’s conversion efficiency is too low or mortality too high to exploit the pest population, even with food supplements. Lemma 4.3 The axial equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_2 = \left( 0, \frac{\delta \xi - m (1+\alpha \xi )}{\epsilon (1 + \alpha \xi )}\right)$$\end{document} exists in positive xy-quadrant and is saddle if\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \delta \xi - m (1+\alpha \xi )>\ 0. \end{aligned}$$\end{document} The characteristic equation corresponding to the interior equilibria is given as5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \lambda ^2 - \text {tr } J \bigg |_{(x^*,y^*)} \lambda + \text {Det } J \bigg |_{(x^*,y^*)} = 0. \end{aligned}$$\end{document}where,6\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \begin{aligned} \text {Det } J \bigg |_{(x^*,y^*)} =&\ \frac{2 \delta x^3 y (1 +\alpha \xi - \xi )}{(1+x^2+\alpha \xi )^3} + \epsilon x y \left( \frac{1}{\gamma } + \frac{(1+\alpha \xi - x^2) y}{(1+\alpha \xi + x^2)^2}\right) \bigg |_{(x^*,y^*)}, \end{aligned} \end{aligned}$$\end{document}and7\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \begin{aligned} \text {tr } J \bigg |_{(x^*,y^*)}&= \left( - \epsilon y - \frac{x}{\gamma } + \frac{xy (x^2 - 1 - \alpha \xi )}{(1+x^2 + \alpha \xi )^2} \right) \bigg |_{(x^*,y^*)}. \end{aligned} \end{aligned}$$\end{document}From (6) and (7), we can conclude that the determinent of the jacobian is positive when \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$1 + \alpha \xi - \xi> 0$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${x^*}^2 < 1 + \alpha \xi$$\end{document}. The trace of the jacobian is negative when \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${x^*}^2 < 1 + \alpha \xi$$\end{document}. For the interior equilibrium, stability conditions depend on the relationship among food parameters, competition strength, and conversion efficiency. When these satisfy particular inequalities, the coexistence state becomes asymptotically stable, ensuring sustained prey-predator coexistence. Otherwise, the equilibrium becomes unstable, leading to oscillations or extinction of one species. Theorem 4.4 The interior equilibrium of the system (2), if it exists, is asymptotically stable when \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$1 + \alpha \xi - \xi> 0$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{\epsilon }{1+\frac{\epsilon }{\gamma }}< x^* < \text {min } \{ \sqrt{1 + \alpha \xi }, \gamma \}$$\end{document}.saddle otherwise. asymptotically stable when \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$1 + \alpha \xi - \xi> 0$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{\epsilon }{1+\frac{\epsilon }{\gamma }}< x^* < \text {min } \{ \sqrt{1 + \alpha \xi }, \gamma \}$$\end{document}. asymptotically stable when \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$1 + \alpha \xi - \xi> 0$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{\epsilon }{1+\frac{\epsilon }{\gamma }}< x^* < \text {min } \{ \sqrt{1 + \alpha \xi }, \gamma \}$$\end{document}. saddle otherwise. saddle otherwise. Remark 3 (Ecological interpretation) The stability condition \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$x^{*} < \sqrt{1+\alpha \xi }$$\end{document} has a clear biological meaning. The term \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$1+\alpha \xi$$\end{document} represents the effective half-saturation level modified by additional food. When the equilibrium pest density is below this level, the predator’s functional response is in its accelerating (convex) phase, meaning any increase in pest density leads to a more-than-proportional increase in predation. This negative feedback stabilises the coexistence equilibrium. Conversely, when \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$x^{*}> \sqrt{1+\alpha \xi }$$\end{document}, the response is in its decelerating (concave) phase, and predation cannot keep pace with pest growth, leading to instability. Biologically, this implies that excessive or low-quality supplemental feeding can destabilize the system, while moderate provisioning promotes persistence of both populations. Bifurcation analysis Population systems often undergo qualitative transitions as environmental parameters vary. In this context, bifurcation theory helps identify parameter thresholds where system behaviour changes abruptly. In this section, we study the various possible bifurcations exhibited by the system (2). We prove the transcritical, saddle-node and Hopf bifurcations analytically. Further, we numerically simulate Hopf bifurcation, Focus-node transitions and Hysterisis loop in addition to these two bifurcations. Transcritical bifurcation A transcritical bifurcation occurs when the stability of the prey-only and coexistence equilibria is exchanged as the additional-food quantity crosses a critical value. This transition marks the boundary between predator extinction and stable coexistence. Theorem 5.1 Let \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi ^{*} = \frac{m(1+\gamma ^{2}) - \delta \gamma ^{2}}{\delta - m\alpha }$$\end{document} and suppose the following non-degeneracy conditions hold: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta \ne m\alpha$$\end{document} (ensures \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi ^{*}$$\end{document} is well-defined),\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$1 + (1-\alpha )\gamma ^{2} \ne 0$$\end{document} (ensures the transversality condition is non-zero),\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi ^{*}> 0$$\end{document} (ensures the bifurcation occurs for a physically meaningful food quantity). \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta \ne m\alpha$$\end{document} (ensures \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi ^{*}$$\end{document} is well-defined), \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta \ne m\alpha$$\end{document} (ensures \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi ^{*}$$\end{document} is well-defined), \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$1 + (1-\alpha )\gamma ^{2} \ne 0$$\end{document} (ensures the transversality condition is non-zero), \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$1 + (1-\alpha )\gamma ^{2} \ne 0$$\end{document} (ensures the transversality condition is non-zero), \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi ^{*}> 0$$\end{document} (ensures the bifurcation occurs for a physically meaningful food quantity). \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi ^{*}> 0$$\end{document} (ensures the bifurcation occurs for a physically meaningful food quantity). Then the system (2) undergoes a transcritical bifurcation at the axial equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1 = (\gamma , 0)$$\end{document} when \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi = \xi ^{*}$$\end{document}. Remark 4 Condition (a) requires that additional food quality α differs from δ /m; otherwise, varying ξ does not change the stability of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1$$\end{document}. Condition (b) requires \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\gamma \ne 1/\sqrt{\alpha - 1}$$\end{document} when α> 1; ecologically, this excludes a degenerate case where the prey carrying capacity exactly compensates the food quality effect. Condition (c) ensures \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi ^{*}> 0$$\end{document}, which requires either \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta> m\alpha$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta \gamma ^{2}> m(1+\gamma ^{2})$$\end{document}, or \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta < m\alpha$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta \gamma ^{2} < m(1+\gamma ^{2})$$\end{document}. Proof The jacobian matrix corresponding to equilibrium point \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1=(\gamma ,0)$$\end{document} is given by\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} J(E_1) = \begin{bmatrix} -1 & \frac{-\gamma ^2}{1 + \alpha \xi + \gamma ^2} \\ \ 0 & \frac{(\delta - m) \gamma ^2 + \delta \xi - m (1+\alpha \xi )}{1 + \alpha \xi + \gamma ^2} \end{bmatrix}. \end{aligned}$$\end{document}The eigenvectors corresponding to the zero eigenvalues of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$J(E_1)$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$J(E_1)^{T}$$\end{document} be denoted by V and W, respectively.\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} V = \begin{bmatrix} V_1 \\ V_2 \end{bmatrix} = \begin{bmatrix} 1 \\ - \frac{1 + \alpha \xi + \gamma ^2 }{\gamma ^2} \end{bmatrix} , \ \ W = \begin{bmatrix} 0 \\ 1 \end{bmatrix}. \end{aligned}$$\end{document}Note that \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$V_2 < 0$$\end{document}. Let us denote system (2) as \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$H = \begin{bmatrix} F \\ G \end{bmatrix}.$$\end{document} Thus,\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} H_{\xi } (E_1; \xi ^*)= & \begin{bmatrix} 0 \\ 0 \end{bmatrix}, \\ DH_\xi (E_1; \xi ^*)V= & \begin{bmatrix} \frac{\partial F_\xi }{\partial x} & \frac{\partial F_\xi }{\partial y} \\ \frac{\partial G_\xi }{\partial x} & \frac{\partial G_\xi }{\partial y} \end{bmatrix} \begin{bmatrix} V_1 \\ V_2 \end{bmatrix} _{(E_1; \xi ^*)} \\= & \frac{-1}{1+\alpha \xi + \gamma ^2} \begin{bmatrix} \alpha \\ \frac{\delta }{\gamma ^2} \left( 1 + (1-\alpha )\gamma ^2 \right) \end{bmatrix}, \\ D^2 H(E_1; \xi ^*)(V, V)= & \begin{bmatrix} \frac{\partial ^2 F}{\partial x^2} V_1^2 + 2 \frac{\partial ^2 F}{\partial x \partial y} V_1 V_2 + \frac{\partial ^2 F}{\partial y^2} V_2^2 \\ \frac{\partial ^2 G}{\partial x^2} V_1^2 + 2 \frac{\partial ^2 G}{\partial x \partial y} V_1 V_2 + \frac{\partial ^2 G}{\partial y^2} V_2^2 \end{bmatrix} _{(E_1, \xi ^*)} \\= & \begin{bmatrix} \frac{2 (1 + \alpha \xi - \gamma ^2)}{ \gamma (1 + \alpha \xi + \gamma ^2)} \\ \frac{-2 \epsilon (1 + \alpha \xi + \gamma ^2)^3 - 4 \delta \gamma ^3 (1+\alpha \xi - \xi )}{\gamma ^4 (1 + \alpha \xi + \gamma ^2)} \end{bmatrix}. \end{aligned}$$\end{document}If \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$1+ (1-\alpha )\gamma ^2 \ne 0$$\end{document}, we have\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} W^{T} H_{\xi } (E_1; \xi ^*)= & 0, \\ W^{T} [DH_\xi (E_1; \xi ^*)V]= & \frac{-\delta (1+(1-\alpha )\gamma ^2)}{\gamma ^2 (1 + \alpha \xi + \gamma ^2)} \ne 0, \\ W^{T} [D^2 H(E_1; \xi ^*)(V, V) ]= & -2 \epsilon (1 + \alpha \xi + \gamma ^2)^3 - 4 \delta \gamma ^3 (1+\alpha \xi - \xi ) \\ & \ne 0. \end{aligned}$$\end{document}By the Sotomayor’s theorem23, the system (2) undergoes a transcritical bifurcation around \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1$$\end{document} at \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi = \xi ^*$$\end{document}. □ Remark 5 (Verification via Perko’s framework) The conditions of Sotomayor’s theorem, as stated in Perko23, Theorem 1, Section 4.6, require: (S1) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$W^{T}H_{\xi }(E_1;\xi ^{*}) = 0$$\end{document}; (S2) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$W^{T}[DH_{\xi }(E_1;\xi ^{*})V] \ne 0$$\end{document} (transversality); (S3) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$W^{T}[D^{2}H(E_1;\xi ^{*})(V,V)] \ne 0$$\end{document} (non-degeneracy). All three conditions are verified analytically in the proof above. For the parameter set γ = 15, α = 0.1, δ = 0.45, m = 0.28, we compute \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi ^{*} = \frac{0.28(1+225) - 0.45 \times 225}{0.45 - 0.28 \times 0.1} = 2.0238$$\end{document}, and verify numerically:(S2): \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$W^{T}[DH_{\xi }V] = -0.0042 \ne 0$$\end{document},(S3): \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$W^{T}[D^{2}H(V,V)] = -1.274 \ne 0$$\end{document}. (S2): \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$W^{T}[DH_{\xi }V] = -0.0042 \ne 0$$\end{document}, (S2): \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$W^{T}[DH_{\xi }V] = -0.0042 \ne 0$$\end{document}, (S3): \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$W^{T}[D^{2}H(V,V)] = -1.274 \ne 0$$\end{document}. (S3): \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$W^{T}[D^{2}H(V,V)] = -1.274 \ne 0$$\end{document}. Similarly, for the saddle-node bifurcation at \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_2$$\end{document} (Theorem 5.2), \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi ^{*} = m/(\delta - m\alpha )$$\end{document} and the conditions of Perko23, Theorem 2, Section 4.6 are verified: (SN1) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$W^{T}H_{\xi } \ne 0$$\end{document}, (SN2) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$W^{T}[D^{2}H(V,V)] = -2\epsilon \ne 0$$\end{document}. The transcritical and saddle node behaviors are numerically depicted in Figure 2 and Figure 3 respectively.Fig. 2Transcritical bifurcation diagram around axial equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1 = (\gamma ,0)$$\end{document} with respect to the quantity of additional food ξ. Transcritical bifurcation diagram around axial equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1 = (\gamma ,0)$$\end{document} with respect to the quantity of additional food ξ. Saddle-node bifurcation A saddle-node bifurcation arises when two equilibria merge and annihilate each other, producing abrupt shifts in population states. The mathematical condition for its occurrence depends on the same control parameters, particularly the quantity and quality of additional food. Theorem 5.2 When the parameter satisfies \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta \ne m \alpha$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi = \xi ^* = \frac{m}{\delta - m \alpha }$$\end{document}, a saddle-node bifurcation occurs at \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_2 = \left( 0, \frac{\delta \xi - m (1+\alpha \xi )}{\epsilon (1 + \alpha \xi )} \right)$$\end{document} in the system (2). Proof The jacobian matrix corresponding to equilibrium point \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_2 = \left( 0, \frac{\delta \xi - m (1+\alpha \xi )}{\epsilon (1+\alpha \xi )} \right)$$\end{document} is given by\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} J(E_2) = \begin{bmatrix} 1 & 0 \\ 0 & - \frac{\delta \xi - m (1 + \alpha \xi )}{1 + \alpha \xi } \end{bmatrix}. \end{aligned}$$\end{document}The eigenvectors corresponding to the zero eigenvalues of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$J(E_2)$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$J(E_2)^{T}$$\end{document} be denoted by V and W, respectively.\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} V = \begin{bmatrix} V_1 \\ V_2 \end{bmatrix} = \begin{bmatrix} 0 \\ 1 \end{bmatrix}, \ \ W = \begin{bmatrix} 0 \\ 1 \end{bmatrix}. \end{aligned}$$\end{document}Let us denote system (2) as \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$H = \left[ \begin{matrix} F \\ G \end{matrix} \right] .$$\end{document} Thus,\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} H_{\xi } (E_2; \xi ^*)= & \begin{bmatrix} 0 \\ \frac{\delta (\delta \xi - m (1+\alpha \xi ))}{\epsilon (1 + \alpha \xi )^3} \end{bmatrix}, \\ DH_\xi (E_2; \xi ^*)V= & \begin{bmatrix} \frac{\partial F_\xi }{\partial x} & \frac{\partial F_\xi }{\partial y} \\ \frac{\partial G_\xi }{\partial x} & \frac{\partial G_\xi }{\partial y} \end{bmatrix} \begin{bmatrix} V_1 \\ V_2 \end{bmatrix} _{(E_2; \xi ^*)} = \begin{bmatrix} 0 \\ \frac{\delta }{(1 + \alpha \xi )^2} \end{bmatrix}, \\ D^2 H(E_2; \xi ^*)(V, V)= & \begin{bmatrix} \frac{\partial ^2 F}{\partial x^2} V_1^2 + 2 \frac{\partial ^2 F}{\partial x \partial y} V_1 V_2 + \frac{\partial ^2 F}{\partial y^2} V_2^2 \\ \frac{\partial ^2 G}{\partial x^2} V_1^2 + 2 \frac{\partial ^2 G}{\partial x \partial y} V_1 V_2 + \frac{\partial ^2 G}{\partial y^2} V_2^2 \end{bmatrix}_{(E_2, \xi ^*)} = \begin{bmatrix} 0 \\ - 2 \epsilon \end{bmatrix}. \end{aligned}$$\end{document}If \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta \ne m \alpha$$\end{document}, we have\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} W^{T} H_{\xi } (E_2; \xi ^*)= & \frac{\delta (\delta \xi - m (1+\alpha \xi ))}{\epsilon (1 + \alpha \xi )^3} \ne 0, \\ W^{T} [DH_\xi (E_2; \xi ^*)V]= & \frac{\delta }{(1 + \alpha \xi )^2} \ne 0, \\ W^{T} [D^2 H(E_2; \xi ^*)(V, V) ]= & - 2 \epsilon \ne 0. \end{aligned}$$\end{document}By the Sotomayor’s theorem23, the system (2) undergoes a saddle-node bifurcation around \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_2$$\end{document} at \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi = \xi ^*$$\end{document}. □ This behavior is numerically depicted in Figure 3.Fig. 3Saddle-node bifurcation diagram around axial equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_2 = \left( 0,\frac{\delta \xi - m (1 + \alpha \xi )}{\epsilon (1+\alpha \xi )}\right)$$\end{document} with respect to the quantity of additional food ξ. Saddle-node bifurcation diagram around axial equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_2 = \left( 0,\frac{\delta \xi - m (1 + \alpha \xi )}{\epsilon (1+\alpha \xi )}\right)$$\end{document} with respect to the quantity of additional food ξ. This Figure shows the bifurcation diagram of interior equilibria as the additional food quantity ξ varies, revealing two saddle-node bifurcation points at SN1 (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi \approx 3.69$$\end{document}) and SN2 (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\xi \approx 8.05$$\end{document}) where stable and saddle branches collide and annihilate. The left panel tracks the prey equilibrium x* and the right panel tracks the predator equilibrium y*, both exhibiting the characteristic S-shaped curve with stable branches (blue) connected by an unstable saddle branch (red). The region between SN1 and SN2 admits bistability - two stable coexistence states coexist - implying that the long-term pest level depends on initial conditions and the history of food supplementation. Hopf bifurcation The interior equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E^{*} = (x^{*}, y^{*})$$\end{document} undergoes a Hopf bifurcation when the trace of the Jacobian vanishes while the determinant remains positive. From equations (7) and (6), we have: Theorem 5.3 Let \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E^{*} = (x^{*}, y^{*})$$\end{document} be an interior equilibrium of system (2) with \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\textrm{Det}\,J|_{E^{*}}> 0$$\end{document}. Suppose there exists a critical value \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon = \epsilon _{H}$$\end{document} such that: (i)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\textrm{tr}\,J|_{E^{*}}(\epsilon _{H}) = 0$$\end{document}, i.e., 8\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \epsilon _{H}\,y^{*} + \frac{x^{*}}{\gamma } = \frac{x^{*}y^{*}\bigl ((x^{*})^{2} - 1 - \alpha \xi \bigr )}{\bigl (1 + (x^{*})^{2} + \alpha \xi \bigr )^{2}}; \end{aligned}$$\end{document}(ii)The transversality condition holds: 9\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \frac{d}{d\epsilon }\,\textrm{tr}\,J\Big |_{\epsilon = \epsilon _{H}} \ne 0. \end{aligned}$$\end{document} (i)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\textrm{tr}\,J|_{E^{*}}(\epsilon _{H}) = 0$$\end{document}, i.e., 8\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \epsilon _{H}\,y^{*} + \frac{x^{*}}{\gamma } = \frac{x^{*}y^{*}\bigl ((x^{*})^{2} - 1 - \alpha \xi \bigr )}{\bigl (1 + (x^{*})^{2} + \alpha \xi \bigr )^{2}}; \end{aligned}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\textrm{tr}\,J|_{E^{*}}(\epsilon _{H}) = 0$$\end{document}, i.e., 8\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \epsilon _{H}\,y^{*} + \frac{x^{*}}{\gamma } = \frac{x^{*}y^{*}\bigl ((x^{*})^{2} - 1 - \alpha \xi \bigr )}{\bigl (1 + (x^{*})^{2} + \alpha \xi \bigr )^{2}}; \end{aligned}$$\end{document} (ii)The transversality condition holds: 9\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \frac{d}{d\epsilon }\,\textrm{tr}\,J\Big |_{\epsilon = \epsilon _{H}} \ne 0. \end{aligned}$$\end{document} The transversality condition holds: 9\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \frac{d}{d\epsilon }\,\textrm{tr}\,J\Big |_{\epsilon = \epsilon _{H}} \ne 0. \end{aligned}$$\end{document} Then the system undergoes a Hopf bifurcation at \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E^{*}$$\end{document} when \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon = \epsilon _{H}$$\end{document}. Proof At \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon = \epsilon _{H}$$\end{document}, the characteristic equation becomes \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\lambda ^{2} + \textrm{Det}\,J|_{E^{*}} = 0$$\end{document}, yielding purely imaginary eigenvalues \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\lambda _{1,2} = \pm i\sqrt{\textrm{Det}\,J|_{E^{*}}}$$\end{document}. The transversality condition (9) ensures that the eigenvalues cross the imaginary axis with non-zero speed as ε varies through \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon _{H}$$\end{document}. To verify the direction of crossing, we compute\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \frac{d}{d\epsilon }\,\textrm{tr}\,J\Big |_{E^{*}} = -y^{*} + \text {(implicit terms from } \frac{dx^{*}}{d\epsilon },\, \frac{dy^{*}}{d\epsilon }\text {)}. \end{aligned}$$\end{document}The dominant contribution is \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$-y^{*} < 0$$\end{document}, indicating that as ε increases through \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon _{H}$$\end{document}, the trace decreases through zero, and stability is regained. This is consistent with a supercritical Hopf bifurcation, where a stable limit cycle exists for \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon < \epsilon _{H}$$\end{document} and shrinks to a point at \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon _{H}$$\end{document}. □ Remark 6 (First Lyapunov coefficient) The sign of the first Lyapunov coefficient \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\ell _{1}$$\end{document} determines whether the Hopf bifurcation is supercritical (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\ell _{1} < 0$$\end{document}) or subcritical (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\ell _{1}> 0$$\end{document}). For the parameter set γ = 15, α = 0.1, ξ = 0.45, δ = 0.45, m = 0.28, numerical computation yields \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\ell _{1} \approx -0.0037 < 0$$\end{document}, confirming a supercritical Hopf bifurcation. Numerical simulations also reveal Hopf bifurcations, where a stable equilibrium loses stability and periodic oscillations emerge. As intra-specific competition or other parameters vary, these oscillations may disappear again, restoring equilibrium through reverse bifurcation. The existence of Hopf bifurcation with respect to ε is depicted in Figure 4. Stable limit cycle is observed around the interior equilibrium when ε = 0.035 and the limit cycle disappears when ε is increased to 0.045. The remaining parameter values are as follows: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\gamma = 15.0,\ \alpha = 0.1,\ \xi = 0.45,\ \delta = 0.45,\ m=0.28$$\end{document}.Fig. 4Supercritical Hopf bifurcation with respect to the intra-specific competition parameter ε. Left panel: at ε = 0.035, a stable limit cycle (yellow trajectory) surrounds an unstable focus, indicating persistent oscillations in pest and predator populations. Right panel: increasing competition to ε = 0.045 stabilises the interior equilibrium (red dot), eliminating oscillations. This demonstrates that stronger intra-specific competition among predators can dampen population cycles and promote steady coexistence. Parameters: γ = 15.0, α = 0.1, ξ = 0.45, δ = 0.45, m = 0.28. Supercritical Hopf bifurcation with respect to the intra-specific competition parameter ε. Left panel: at ε = 0.035, a stable limit cycle (yellow trajectory) surrounds an unstable focus, indicating persistent oscillations in pest and predator populations. Right panel: increasing competition to ε = 0.045 stabilises the interior equilibrium (red dot), eliminating oscillations. This demonstrates that stronger intra-specific competition among predators can dampen population cycles and promote steady coexistence. Parameters: γ = 15.0, α = 0.1, ξ = 0.45, δ = 0.45, m = 0.28. Hysteresis loop The interplay of these bifurcations gives rise to hysteresis phenomena–situations where the final state depends on the history of parameter change. Ecologically, this means that gradual adjustments in supplemental feeding can lead to sudden and irreversible shifts between low- and high-pest states, even under small environmental variations. The interior equilibria corresponding to different values of ε are illustrated in Figure 5. The subplots display the intra-specific population plotted against the interior prey and predator equilibrium values, respectively. The parameter values used for this numerical simulation are \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\gamma = 15.0, \ \alpha = 0.1, \ \xi = 1.0,\ \delta = 0.3,\ m= 0.258$$\end{document}. This Figure demonstrates the presence of multiple interior equilibria across the parameter range \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon \in [0.002, 0.02]$$\end{document}, highlighting the system’s nonlinear response and the emergence of potential bistable regimes within this interval. This S-shaped plot also depicts the existence of two saddle-node bifurcations at the two-turning points (folds) of the S-curves in both the panels. Also two focus-node transitions are observed in these plots. Though eigenvalue types are changing from real to complex, these are not bifurcations in the topological sense. In addition to this, a change in stable focus to unstable focus is observed in this plot. However, this doesn’t guarantee the existence or disappearance of limit cycle. Therefore, this stability transition of a fixed point need not be a Hopf bifurcation.Fig. 5Nature of interior equilibria with respect to the intra-specific competition (ε). Nature of interior equilibria with respect to the intra-specific competition (ε). Figure 5 presents two stable equilibria branches. Let us refer the bottom and top stable branches in this S-shaped curve as the branch 1 and 2 respectively. As ε increases, stability remains in the branch 1 until the second saddle-node bifurcation point is reached. Then there will be a jump in the stability to the branch 2. Similarly, as ε reduces, the system continues to remain stable in branch 2 until it jumps back at the other fold. Therefore, the history of ε determines the stable equilibria in the bistable behavior. Now, when we consider a time-varying ε in the region, we observe the hysteresis loop. In Figure 6, we consider the ε as smooth periodic sweep using sine in the region (0.002, 0.02). Figure 6 represents the hysteresis loop present in the plots with prey interior equilibria, predator interior equilibria, phase portrait respectively. They go up following one branch and come down following different branch. They intersect at the two saddle-node bifurcation points resulting in a loop known as Hysteresis loop (Table 2).Fig. 6Hysteresis loop in the prey–predator system under time-varying intra-specific competition \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon (t) = 0.011 + 0.009\sin (2\pi t/T)$$\end{document}, T = 500. Top panels show that the prey and predator equilibria follow different paths as ε increases vs. decreases, with abrupt jumps at the two saddle-node bifurcation points. Bottom panels show the corresponding phase-plane trajectory forming a closed loop. This hysteresis implies that a temporary increase in predator competition (e.g., due to crowding) can cause an irreversible shift to a high-pest state. Parameters: γ = 15.0, α = 0.1, ξ = 1.0, δ = 0.3, m = 0.258. Hysteresis loop in the prey–predator system under time-varying intra-specific competition \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon (t) = 0.011 + 0.009\sin (2\pi t/T)$$\end{document}, T = 500. Top panels show that the prey and predator equilibria follow different paths as ε increases vs. decreases, with abrupt jumps at the two saddle-node bifurcation points. Bottom panels show the corresponding phase-plane trajectory forming a closed loop. This hysteresis implies that a temporary increase in predator competition (e.g., due to crowding) can cause an irreversible shift to a high-pest state. Parameters: γ = 15.0, α = 0.1, ξ = 1.0, δ = 0.3, m = 0.258. The focus-node transitions observed in Figure 5 correspond to changes in the discriminant \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(\textrm{tr}\,J)^{2} - 4\,\textrm{Det}\,J$$\end{document} of the characteristic equation. When this discriminant changes sign, eigenvalues transition between real (node) and complex (focus) types. As noted by Strogatz24 and Kuznetsov25, such transitions are smooth and do not constitute topological bifurcations—the phase portrait does not change qualitatively. The hysteresis phenomenon follows from the coexistence of two stable equilibria separated by a saddle in the bistability region, a standard feature of systems with two saddle-node bifurcations on the same branch; see Strogatz24, Chapter 3.6.Table 2Parameter values used in numerical simulations and their sources.FigureParametersValuesSource/RationaleFig. 2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\gamma , \alpha , \delta , m, \epsilon$$\end{document}15, 0.1, 0.45, 0.28, 0.035Chosen to place the system near thetranscritical bifurcation threshold;\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\gamma = K/a \approx 100/11.83$$\end{document} rounded for clarityFig. 4ε0.035, 0.045Values straddling the numerically detectedHopf bifurcation point \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon _H \approx 0.04$$\end{document}Fig. 5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\gamma , \alpha , \xi , \delta , m$$\end{document}15, 0.1, 1.0, 0.3, 0.258Selected to exhibit an S-shaped equilibriumcurve with bistability in \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon \in [0.002, 0.02]$$\end{document}Fig. 6Same as Fig. 5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon (t) \in [0.002, 0.02]$$\end{document}Periodic sweep to demonstrate hysteresisFigs. 7, 8, 9, 10VariedSee captionSystematic exploration of α–ξ spaceFigs. 11, 12, 13, 14, 15c, a0.967, 11.83Experimentally estimated from19 Table 2Parameter values used in numerical simulations and their sources.FigureParametersValuesSource/RationaleFig. 2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\gamma , \alpha , \delta , m, \epsilon$$\end{document}15, 0.1, 0.45, 0.28, 0.035Chosen to place the system near thetranscritical bifurcation threshold;\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\gamma = K/a \approx 100/11.83$$\end{document} rounded for clarityFig. 4ε0.035, 0.045Values straddling the numerically detectedHopf bifurcation point \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon _H \approx 0.04$$\end{document}Fig. 5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\gamma , \alpha , \xi , \delta , m$$\end{document}15, 0.1, 1.0, 0.3, 0.258Selected to exhibit an S-shaped equilibriumcurve with bistability in \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon \in [0.002, 0.02]$$\end{document}Fig. 6Same as Fig. 5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon (t) \in [0.002, 0.02]$$\end{document}Periodic sweep to demonstrate hysteresisFigs. 7, 8, 9, 10VariedSee captionSystematic exploration of α–ξ spaceFigs. 11, 12, 13, 14, 15c, a0.967, 11.83Experimentally estimated from19 Parameter values used in numerical simulations and their sources. Remark 7 For the dimensionless simulations (Figures 4, 5, 6, 7, 8, 9, 10), parameter values were selected to clearly illustrate specific dynamical phenomena (bifurcations, bistability, hysteresis) rather than to match a particular biological system. The application-oriented simulations (Figures 11, 12, 13, 14, 15) use experimentally derived values where available and sweep assumed parameters over biologically plausible ranges. This two-tier approach separates the general mathematical analysis from the species-specific application. Global dynamics To explore broader parameter interactions, the system (2) is analysed in the two-dimensional space defined by food quality and quantity (α-ξ parameter space). Distinct regions correspond to qualitatively different dynamical behaviours, such as extinction, stable coexistence, or bistability. In the absence of additional food, the system behaves qualitatively as depicted in these three regions of Figure 7.Fig. 7Dynamics of the system (2) in the absence of additional food. Dynamics of the system (2) in the absence of additional food. Now, in each of these regions, we study the influence of additional food by dividing the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\alpha - \xi$$\end{document} parameter space into regions based on the following curves. Each of these curves divide the space into two regions based on the qualitative nature of the equilibrium point.\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} \begin{aligned} \text {Bifurcation Curve for } E_0:&\ \phi _1(\alpha ,\xi ) : \ \delta \xi - m (1 + \alpha \xi ) = 0. \\ \text {Bifurcation Curve for } E_1:&\ \phi _2(\alpha ,\xi ) : \ \delta \xi - m (1 + \alpha \xi ) + (\delta - m) \gamma ^2 = 0. \\ \text {Existence Curve for } E^*:&\ \phi _3(\alpha ,\xi ) : \ 1 + \alpha \xi - \xi = 0. \\ \end{aligned} \end{aligned}$$\end{document}Fig. 8Influence of additional food on the system (2) when the parameters belong to the region \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$R_1$$\end{document}. Influence of additional food on the system (2) when the parameters belong to the region \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$R_1$$\end{document}. Figure 8 divides the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\alpha - \xi$$\end{document} space into 4 regions. In this region, there is no interior equilibrium in the absence of additional food. As additional food is provided, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_0$$\end{document} is unstable and the rest are saddle in nature in the region \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{11}$$\end{document}. When we move to the region \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{12}$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_2$$\end{document} ceases to exist and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E^*$$\end{document} will be the only stable equilibrium. By further increasing the additional food quality and quantity, we will get a bistability condition in the regions \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{13}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{14}$$\end{document} as \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E^*$$\end{document} are stable. However, provision of additional food in the regions \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{13}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{14}$$\end{document} takes us to the original qualitative behavior as in the case of absence of additional food based on the initial prey and predator populations. Phase-plane studies reveal that increasing additional food initially stabilizes the system but may later induce multistability if competition is weak. Maps of these regimes show how ecological balance can be preserved only within a limited parameter window. The analysis further demonstrates that bistability leads to dependence on initial conditions: small differences in starting prey or predator densities may direct the system toward drastically different equilibria. Therefore, controlled management of auxiliary feeding and predator release is crucial to maintain populations within the desired stable regime.Fig. 9Influence of additional food on the system (2) when the parameters belong to the region \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$R_2$$\end{document}.Fig. 10Influence of additional food on the system (2) when the parameters belong to the region \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$R_3$$\end{document}. Influence of additional food on the system (2) when the parameters belong to the region \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$R_2$$\end{document}. Influence of additional food on the system (2) when the parameters belong to the region \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$R_3$$\end{document}. Figures 9 and 10 divides the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\alpha - \xi$$\end{document} space into 5 regions each. In these regions, there is a stable interior equilibrium in the absence of additional food. As additional food is provided, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1,\ E_2,\ E^*$$\end{document} are saddle in the regions \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{21}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{31}$$\end{document}. The interior equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E^*$$\end{document} and the predator-free equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1$$\end{document} are stable in regions \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{25}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{35}$$\end{document} resulting in bistability of the system. In rest of the regions, only interior equilibrium is stable. Therefore, as additional food is provided the system continues to exhibit the original nature of initial system depending on the initial prey and predator populations. Provision of additional food - consequences In this section, we present the consequences of providing additional food by studying the possibility of existence in three different scenarios: Pest eradication, pest dominance and the coexistence of pest and natural enemies. The ecological implications of supplementary feeding are multifaceted. Although intended to support predator populations, indiscriminate provisioning can undermine pest suppression. When excessive, it may reduce predation pressure on pests, allowing their resurgence; when insufficient, predators fail to survive. Three qualitative outcomes can be recognized. First, complete pest eradication is theoretically impossible under the given functional response, as the pest-free equilibrium is never globally stable. Second, dominance of the pest species arises only under specific bistable conditions where predators fail to establish. Third, stable coexistence occurs when parameters favour balanced growth and competition. As pest-free equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_2$$\end{document} is saddle whenever it exists, the arbitrary provision of additional food can never lead to the complete pest eradication for this system irrespective of the initial behaviour of the system in the absence of additional food. Now consider the case of pest dominance. In this system, there is no clear case of pest dominance possible as the natural enemy free equilibrium is stable only in the regions \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{13}, \ A_{14},\ A_{25}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{35}$$\end{document}. However in these regions, the system exhibits bistability where both the equilibria \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E^*$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1$$\end{document} are stable. The dynamics goes to one of these equilibria depending on the initial condition. Therefore, altering the natural enemy population can change the stability of the system and avoid pest dominance stage. In the rest of the regions, only interior equilibria \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E^*$$\end{document} is stable. As the system can exhibit upto 5 interior equilibria and the real positive equilibria are either saddle or stable, the system can exhibit bistability among the interior equilibria which highlights the importance of initial prey and predator populations. Therefore, choosing the initial populations so that the system lead to the interior equilibria with least pest population will be the best strategy to reduce pest dominance. Also it is important to note that the minimum pest population that can be reached is \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{\epsilon }{1+\frac{\epsilon }{\gamma }}$$\end{document}. This highlights the importance of intra-specific competition for the pest management. Practical biocontrol guidelines The above mentioned theoretical findings translate into the following actionable recommendations for biological control practitioners: Gradual predator augmentation over mass release. The hysteresis analysis (Section 5.4) shows that the system’s long-term state depends on the path taken through parameter space. A gradual increase in predator numbers keeps the system on the lower (low-pest) branch of the bistability region, whereas a sudden mass release may overshoot the second saddle-node bifurcation, causing an unexpected jump to a high-pest state.Moderate supplementary feeding. The global dynamics analysis (Section 6) reveals that excessive additional food (high ξ, low α) can push the system into bistable regions where the prey-only equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1$$\end{document} becomes stable alongside the coexistence equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E^{*}$$\end{document}. Practitioners should calibrate food quantity to remain in regions \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{12}$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{22}$$\end{document}–\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{24}$$\end{document}, or \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{32}$$\end{document}–\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{34}$$\end{document}, where only the interior equilibrium is stable.Monitor pest-to-predator ratios. The minimum achievable equilibrium pest density is \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{\epsilon }{1+\epsilon /\gamma }$$\end{document} (Theorem 4.4). Since ε = da/c, reducing intra-specific competition (e.g., by ensuring adequate spatial distribution of predators) directly lowers the pest floor.Irreversibility warning. The hysteresis loop (Figure 6) implies that once the system jumps to a high-pest branch, reversing to the low-pest state requires reducing ε (or equivalently, adjusting predator density) well past the original bifurcation point. This time lag between cause and effect must be factored into management schedules. Gradual predator augmentation over mass release. The hysteresis analysis (Section 5.4) shows that the system’s long-term state depends on the path taken through parameter space. A gradual increase in predator numbers keeps the system on the lower (low-pest) branch of the bistability region, whereas a sudden mass release may overshoot the second saddle-node bifurcation, causing an unexpected jump to a high-pest state. Gradual predator augmentation over mass release. The hysteresis analysis (Section 5.4) shows that the system’s long-term state depends on the path taken through parameter space. A gradual increase in predator numbers keeps the system on the lower (low-pest) branch of the bistability region, whereas a sudden mass release may overshoot the second saddle-node bifurcation, causing an unexpected jump to a high-pest state. Moderate supplementary feeding. The global dynamics analysis (Section 6) reveals that excessive additional food (high ξ, low α) can push the system into bistable regions where the prey-only equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1$$\end{document} becomes stable alongside the coexistence equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E^{*}$$\end{document}. Practitioners should calibrate food quantity to remain in regions \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{12}$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{22}$$\end{document}–\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{24}$$\end{document}, or \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{32}$$\end{document}–\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{34}$$\end{document}, where only the interior equilibrium is stable. Moderate supplementary feeding. The global dynamics analysis (Section 6) reveals that excessive additional food (high ξ, low α) can push the system into bistable regions where the prey-only equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1$$\end{document} becomes stable alongside the coexistence equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E^{*}$$\end{document}. Practitioners should calibrate food quantity to remain in regions \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{12}$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{22}$$\end{document}–\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{24}$$\end{document}, or \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{32}$$\end{document}–\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{34}$$\end{document}, where only the interior equilibrium is stable. Monitor pest-to-predator ratios. The minimum achievable equilibrium pest density is \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{\epsilon }{1+\epsilon /\gamma }$$\end{document} (Theorem 4.4). Since ε = da/c, reducing intra-specific competition (e.g., by ensuring adequate spatial distribution of predators) directly lowers the pest floor. Monitor pest-to-predator ratios. The minimum achievable equilibrium pest density is \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{\epsilon }{1+\epsilon /\gamma }$$\end{document} (Theorem 4.4). Since ε = da/c, reducing intra-specific competition (e.g., by ensuring adequate spatial distribution of predators) directly lowers the pest floor. Irreversibility warning. The hysteresis loop (Figure 6) implies that once the system jumps to a high-pest branch, reversing to the low-pest state requires reducing ε (or equivalently, adjusting predator density) well past the original bifurcation point. This time lag between cause and effect must be factored into management schedules. Irreversibility warning. The hysteresis loop (Figure 6) implies that once the system jumps to a high-pest branch, reversing to the low-pest state requires reducing ε (or equivalently, adjusting predator density) well past the original bifurcation point. This time lag between cause and effect must be factored into management schedules. Applications to Bemisia Tabaci In Shahpouri et al.19, single female Orius albidipennis were exposed to varying densities (5–35 individuals) of third-instar Bemisia tabaci nymphs on cucumber leaf discs for a 24-hour period. Logistic regression and Roger’s random predator equation were applied to determine the functional response type and estimate predation parameters.Table 3Parameter sources for the O. albidipennis–B. tabaci model.ParameterValueSourceJustificationc\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$0.967\;\text {prey}\cdot \text {pred}^{-1}\cdot \text {h}^{-1}$$\end{document}Shahpouri et al.19Estimated from \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$1/T_h$$\end{document}a11.83 nymphs/discShahpouri et al.19Derived from \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\sqrt{c/b}$$\end{document}r\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$0.05\;\text {h}^{-1}$$\end{document}AssumedTypical whitefly intrinsic growth rate on cucumber at 25\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${}^{\circ }$$\end{document}CK100 nymphs/discAssumedCarrying capacity consistent with leaf disc area (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\approx 20\;\text {cm}^{2}$$\end{document})\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta _1$$\end{document}0.005–0.2 (swept)Assumed rangeConversion efficiencies typical for anthocorid predators\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m_1$$\end{document}0.005–0.05 (swept)Assumed rangeNatural mortality range for Orius spp. in laboratoryd0.001–0.01 (swept)Assumed rangeDensity-dependent mortality reflecting interference competitionα0.1AssumedLow-quality supplement (e.g., pollen vs. prey)η , AVariedAssumedControl variables for food quantity Table 3Parameter sources for the O. albidipennis–B. tabaci model.ParameterValueSourceJustificationc\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$0.967\;\text {prey}\cdot \text {pred}^{-1}\cdot \text {h}^{-1}$$\end{document}Shahpouri et al.19Estimated from \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$1/T_h$$\end{document}a11.83 nymphs/discShahpouri et al.19Derived from \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\sqrt{c/b}$$\end{document}r\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$0.05\;\text {h}^{-1}$$\end{document}AssumedTypical whitefly intrinsic growth rate on cucumber at 25\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${}^{\circ }$$\end{document}CK100 nymphs/discAssumedCarrying capacity consistent with leaf disc area (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\approx 20\;\text {cm}^{2}$$\end{document})\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta _1$$\end{document}0.005–0.2 (swept)Assumed rangeConversion efficiencies typical for anthocorid predators\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m_1$$\end{document}0.005–0.05 (swept)Assumed rangeNatural mortality range for Orius spp. in laboratoryd0.001–0.01 (swept)Assumed rangeDensity-dependent mortality reflecting interference competitionα0.1AssumedLow-quality supplement (e.g., pollen vs. prey)η , AVariedAssumedControl variables for food quantity Parameter sources for the O. albidipennis–B. tabaci model. The sensitivity of model outcomes to these assumed parameters is explored through the heatmaps in Figures 12, 13 and the stability region plots in Figures 14, 15, which sweep broad parameter ranges rather than relying on single point estimates. The experimental validation presented in this section is structured in three tiers. First, the Holling type-III functional response assumption - the core biological hypothesis of the model - is directly validated by fitting the consumption data of Shahpouri et al.19 to the sigmoidal curve, yielding parameters c = 0.967 and a = 11.83 (Fig. 11). Second, the persistence threshold predicted by Lemma 4.2 is shown to be consistent with the sustained predation observed across all tested prey densities in19. Third, the bistability prediction under supplementary feeding constitutes a falsifiable experimental hypothesis, stated explicitly in Table 3. Results indicated a Holling type-III functional response, characterized by slow predation at low prey density, accelerated consumption at intermediate levels, and saturation at high density. The estimated handling time was\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} T_h = 1.034~\text {h per prey}, \end{aligned}$$\end{document}yielding a maximum predation rate\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} c = \frac{1}{T_h} \approx 0.967~\text {prey}\cdot \text {pred}^{-1}\cdot \text {h}^{-1}. \end{aligned}$$\end{document} Parameter estimation For a type-III functional response, the consumption rate at low prey density N can be approximated as\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} C(N) \approx \frac{c}{a^2}N^2, \end{aligned}$$\end{document}where a is the half-saturation constant. Using the empirical coefficient b = 0.0069 reported by Shahpouri et al.19, we obtain\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} a = \sqrt{\frac{c}{b}} = \sqrt{\frac{0.967}{0.0069}} \approx 11.83~\text {nymphs per leaf disc}. \end{aligned}$$\end{document}Thus, the experimentally derived parameters are:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{aligned} c = 0.967, \quad a = 11.83. \end{aligned}$$\end{document}Other parameters of the model \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(r,K,\delta _1,m_1,d,\alpha ,\eta ,A)$$\end{document} are not provided in the paper and must be chosen or fitted from independent data. The model (2) embeds the empirically verified Holling type-III response into a population-dynamic framework. Functional response fit The fitted curve closely reproduces the sigmoidal consumption pattern observed experimentally, with saturation near ∼ 23 prey per day as depicted in Fig. 11.Fig. 11Fitted Holling type-III functional response curve using parameters from Shahpouri et al. (2019). Experimental data points (dots) and theoretical fit (solid line). Fitted Holling type-III functional response curve using parameters from Shahpouri et al. (2019). Experimental data points (dots) and theoretical fit (solid line). Persistence heatmaps With parameters c ≈ 0.967 and a ≈ 11.83 fixed from experimental data, we explore the impact of free variables such as predator mortality (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m_1$$\end{document}), conversion efficiency (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta _1$$\end{document}) and density-dependence (d). Figure 12 simulates the system across a grid of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta _1$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m_1$$\end{document} values and measure predator density at long times. This continuous heatmap displays the long-run predator density after 700 hours of integration across a 65 × 65 grid of conversion efficiency (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta _1$$\end{document}) and predator mortality (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m_1$$\end{document}) values, with intra-specific competition fixed at d = 0.001. The viridis colour gradient progresses from deep purple (near-zero, extinction) in the lower-right corner to yellow-green (high predator density) in the upper-left, revealing a sharp threshold surface that separates persistent from non-persistent predator regimes. The white dashed contour at P = 0.1 nymphs/disc explicitly marks the extinction boundary, confirming analytically that predator establishment requires the ratio \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta _1/m_1$$\end{document} to exceed a critical value - a result fully consistent with the eigenvalue condition for the stability of the prey-only equilibrium \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_1$$\end{document} derived in Lemma 4.2 of the manuscript.Fig. 12Long term predator density for the parameters \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta _1$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m_1$$\end{document}. Long term predator density for the parameters \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta _1$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m_1$$\end{document}. Figure 13 represents the prey suppression heatmap in the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(m_1,d)$$\end{document}-plane. Fixing conversion efficiency at \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta _1 = 0.10$$\end{document}, this plasma-coloured heatmap traces long-run prey density across combinations of predator mortality \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m_1$$\end{document} and intra-specific competition d. High prey density (bright yellow) characterises the right side of the parameter space where predator losses are large, while the left region displays effective suppression well below carrying capacity. The white dashed contour at N = 50 nymphs/disc - precisely half the carrying capacity K = 100 - delineates the boundary of biologically meaningful pest control, and the annotated label makes this threshold immediately readable. Crucially, the near-vertical orientation of the contours indicates that prey suppression is far more sensitive to changes in \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m_1$$\end{document} than to changes in d across the plotted range, a finding with direct practical relevance for biological control programme design.Fig. 13Long term prey density for the parameters d and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m_1$$\end{document}. Long term prey density for the parameters d and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m_1$$\end{document}. Stability regions Simulations across parameter planes \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(\delta _1,m_1)$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(m_1,d)$$\end{document}, depicted in Figs. 14 and 15 respectively, reveal the following.High mortality (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m_1$$\end{document}) or strong density-dependence (d) leads to predator extinction.Intermediate values support stable coexistence.Near the stability boundary, oscillatory dynamics (Hopf bifurcation) arise. High mortality (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m_1$$\end{document}) or strong density-dependence (d) leads to predator extinction. High mortality (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$m_1$$\end{document}) or strong density-dependence (d) leads to predator extinction. Intermediate values support stable coexistence. Intermediate values support stable coexistence. Near the stability boundary, oscillatory dynamics (Hopf bifurcation) arise. Near the stability boundary, oscillatory dynamics (Hopf bifurcation) arise. In Fig. 14, each cell of this 65×65 discrete colour map is assigned to one of four categories - extinction (red), unstable coexistence (orange), stable coexistence (green), or near-Hopf (blue) - based on analytical evaluation of the Jacobian eigenvalue signs at the interior equilibrium. The dominant green region, occupying the upper-left quadrant where conversion efficiency is high and mortality is low, confirms that stable pest-natural enemy coexistence is the generic outcome under favourable ecological conditions. The blue near-Hopf band forms a narrow transitional strip along the stability boundary. Figure 15 is constructed at fixed \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta _1 = 0.10$$\end{document}. This Figure complements Fig. 14 by examining how the interplay between predator mortality and intra-specific competition governs system fate. For these parameter values, the system exhibits only a broad green (stable) region. Unlike Fig. 14, the near-Hopf band is not observed here.Fig. 14Numerical stability regions in \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(\delta _1,m_1)$$\end{document} space. Stable coexistence occurs at high conversion efficiency and low mortality.Fig. 15Numerical stability regions in \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(m_1,d)$$\end{document} space. Numerical stability regions in \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(\delta _1,m_1)$$\end{document} space. Stable coexistence occurs at high conversion efficiency and low mortality. Numerical stability regions in \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(m_1,d)$$\end{document} space. The experimentally validated type-III response enhances ecosystem stability by limiting predation at low prey densities. The model predicts that predator persistence requires moderate mortality and sufficient conversion efficiency. These findings support the use of O. albidipennis as an effective biological control agent under fluctuating prey conditions. Table 4 summarises the correspondence between the key theoretical predictions of the model and the experimental findings from Shahpouri et al.19.Table 4Theoretical predictions vs. experimental evidence for the O. albidipennis–B. tabaci system.Theoretical predictionExperimental evidenceStatusSigmoidal (type-III) functional responseLogistic regression on consumption data yields negative linear and positive quadratic coefficientsConfirmedbegineqnarray*4pt] Predator persistence requires \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta _{1}/m_{1}$$\end{document} above a threshold (Fig. 12)O. albidipennis sustained predation across all tested densities when well-fedConsistentBistability possible under supplementary feedingNot yet tested experimentallyOpen prediction Table 4Theoretical predictions vs. experimental evidence for the O. albidipennis–B. tabaci system.Theoretical predictionExperimental evidenceStatusSigmoidal (type-III) functional responseLogistic regression on consumption data yields negative linear and positive quadratic coefficientsConfirmedbegineqnarray*4pt] Predator persistence requires \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta _{1}/m_{1}$$\end{document} above a threshold (Fig. 12)O. albidipennis sustained predation across all tested densities when well-fedConsistentBistability possible under supplementary feedingNot yet tested experimentallyOpen prediction Theoretical predictions vs. experimental evidence for the O. albidipennis–B. tabaci system. The open prediction regarding bistability represents a testable hypothesis: providing supplementary food (e.g., Ephestia kuehniella eggs) to O. albidipennis colonies at varying levels while monitoring B. tabaci populations should reveal threshold effects consistent with the saddle-node bifurcations predicted in Section 5. In this work, only c and a are directly estimated from the experimental data of19. The remaining parameters are chosen within biologically plausible ranges for anthocorid predators on whitefly hosts. Future experimental work measuring conversion efficiency and density-dependent mortality for O. albidipennis would substantially improve the model’s predictive accuracy. Discussions and conclusions The study provides a comprehensive assessment of a prey–predator system incorporating intra-specific competition and an external food source under a Holling type-III response. The analysis confirms that the system admits bounded solutions and diverse equilibria whose stability depends sensitively on ecological parameters. Bifurcation and hysteresis analyses reveal that multiple equilibria may coexist, and transitions between them can be abrupt. Such behaviour implies that ecological control strategies must avoid parameter regions that induce bistability, as small perturbations could shift the system toward pest dominance. This study comprehensively studied the importance of additional food and intra-specific competition in the dynamics of the system (2). The global dynamics of the system is studied in section 6. In this section, we observed that the provision of additional food can lead to the bistability condition. The system can converge to the pest dominant equilibrium or any of the 5 possible interior equilibria based on the initial conditions. Taking this analysis further, it is observed from the theorem 4.4 that the minimum equilibria pest population is \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{\epsilon }{1+\frac{\epsilon }{\gamma }}$$\end{document}. Therefore, in the case where the stable equilibrium is only any of the interior equilibria, this work gives the least amount of stable pest population, which is directly proportional to the intra-specific competition (ε). As intra-specific competition is due to prey scarcity, higher competition leads to the difficulty in reaching the equilibrium state. Figure 5 depicted the nature of various equilibria with respect to intra-specific competition. This plot depicts the bistability region where the system exhibits two stable interior equilibria. As ε changes with time over the region [0.002, 0.02], the system stabilises either in the top branch or bottom branch of the bistability region depending on the path taken by the trajectory. The dependence on the history of the trajectory plays a crucial role in the biological control of pests. As pest eradication is never stable in this system, keeping the pest at low levels so that they won’t harm the crop, is the next best available strategy. Therefore, the bottom branch in the bistability region is more desirable to the top branch for pest management. Therefore, to remain in the bottom branch, it is advisable to increase natural enemies population over time rather than mass release of natural enemies for this case. Also, the sudden jump after the second saddle-node bifurcation can lead to immediate increase of pests. This sudden surge is drastic not only because of the damage the increased pest can cause to the crop but also the reversibility of damage takes time as the system should reach the first saddle-node bifurcation point. These conclusions provide valuable insights to the biocontrol strategies which is a double-edged sword. The model suggests that optimal management requires careful calibration of both the amount and nutritional quality of supplemental food. Maintaining predators near the coexistence equilibrium ensures minimal pest density while avoiding destabilization. Intra-specific competition plays a protective role by preventing predator overpopulation and the resulting oscillatory outbreaks. Empirical observations confirm that Orius albidipennis follows a Holling type-III functional response in its predation on Bemisia tabaci, consistent with the findings of Shahpouri et al. (2019)19. The estimated model parameters (c = 0.967, a = 11.83) effectively capture the nonlinear feeding dynamics and provide a realistic foundation for theoretical analysis. The validated model successfully reproduces the experimental trends and delineates stability thresholds that align with established ecological predictions. These outcomes highlight the robustness of the Holling type-III framework in describing predator efficiency at varying prey densities and underscore its relevance in designing biocontrol strategies for sustainable pest management. The model adopts a deterministic framework; incorporating environmental and demographic variability in future work could better capture dynamics near critical thresholds. It assumes a well-mixed, spatially uniform system, and can be extended to include spatial heterogeneity, dispersal, and patch dynamics for greater ecological realism. Parameterisation currently relies on limited empirical estimation; further experimental studies could enhance accuracy and predictive strength. The framework focuses on a single prey–predator interaction, with scope to expand toward multi-species systems and more complex ecological relationships. From a practical perspective, the findings highlight that biocontrol is not a linear process. The quality of additional food, competition among predators, and historical population trajectories collectively shape long-term outcomes. Sustainable pest management therefore requires adaptive strategies informed by both theoretical and empirical evidence. Future extensions of this work could incorporate stochastic environmental effects, spatial dispersal, or evolutionary adaptation in predator behaviour. Such additions would enhance the realism of the model and provide deeper insight into how ecological and environmental variability influence biological control success. From a management perspective, the following implications deserve emphasis: Integrated Pest Management (IPM) design. The model’s prediction that pest eradication is never globally stable under a type-III response (Section 6) has a direct IPM implication: practitioners should aim for suppression below economic injury levels rather than eradication. The minimum achievable pest density \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{\epsilon }{1+\epsilon /\gamma }$$\end{document} provides a theoretical lower bound that can guide the setting of action thresholds. Timing of interventions. The hysteresis analysis implies that the order of management actions matters. For instance, first reducing pest density through a targeted intervention (e.g., a compatible reduced-risk pesticide) and then releasing predators positions the system on the low-pest branch, from which it is unlikely to jump to the high-pest state unless competition increases dramatically. Monitoring for early warning. The proximity to saddle-node bifurcation points can be detected through increased variance and autocorrelation in pest population time series—classical early warning signals of critical transitions26. Incorporating such monitoring into biocontrol programmes would allow pre-emptive adjustments to supplementary feeding or predator release rates.