Dynamical Dark Energy or Modified Gravity? Signatures in Gravitational Wave Propagation Purnendu Karmakar, Sandeep Haridasu Abstract We discuss the implications of the most recent DESI Baryon Acoustic Oscillations (BAO) DR2 and DESyr5 compilation of supernovae (SNe) datasets for modified gravity focusing on non-metricity based f ⁡ ( Q ) f(Q) theory, by employing a ‘model-independent’ approach. We reconstruct of dark energy density which is then extended to estimate the perturbation-level quantities, sourced by the modified gravity, namely the effective gravitational coupling, μ \mu , and the amplitude damping parameter of gravitational wave propagation, ν \nu . In light of the remarkable hints for a dynamical dark energy emerging from the analysis of background cosmological data, we discuss the possibility to distinguish between dark energy and modified gravity scenarios from their scalar and tensor perturbation-level signatures. We contrast our findings between the older Pantheon+ and the newer DESyr5 compilations of SNe datasets, which predict a significant distinct signature in the damping parameter of the gravitational wave propagation amplitude. On the other hand, the prediction for the effective gravitational coupling remains insensitive to the choice of the SNe datasets. Our results provide a strong case for the study of gravitational wave propagation in modified gravity theories, in light of the new generation of gravitational wave detectors such as LISA, Einstein Telescope and Cosmic Explorer. I Introduction Alternative geometric reformulations of general relativity Beltrán Jiménez et al. 2019a , particularly those based on the affine connection consisting torsion and non-metricity, have attracted increasing interest in recent years. These approaches provide new perspectives on the nature of gravity comes from geometrical modification. The existence of equivalent formulations of general relativity, such as the teleparallel equivalent (Teleparallel Equivalent of GR) based on liner torsion scalar, and the symmetric teleparallel equivalent (Symmetric Teleparallel Equivalent of GR) based on linear non-metricity scalar Nester and Yo 1999 ; Beltrán Jiménez et al. 2019b , underscores the physical relevance of these extensions and motivates their systematic exploration in the context of modified gravity theories Bahamonde et al. 2023 . One could extend these geometrical models further as a function of scalar curvature, called f(R) model De Felice and Tsujikawa 2010 , torsion called f(T) model Cai et al. 2016 , and non-metricity, called f(T) . Such modifications of gravity allow us explore the more phenomenonlogy of the universe both in the background and perturbation. In this article we take f(Q) as an example, since it recently received decent attentions Barros et al. 2020 ; Ayuso et al. 2021 ; Bajardi et al. 2020 ; Frusciante 2021 ; Atayde and Frusciante 2021 . A model-independent approach to reconstruct f(Q) from observational data is therefore timely. Recent high-precision datasets, such as most recent DESI BAO DRII and Pantheon+, provide an opportunity to infer the shape of the function and the corresponding action, offering added value to the theoretical community. In this work, we address phenomenological testing of f(Q) models and reconstruct their action from current background and perturbation data using symbolic regression and rank the models preferred by the data. Our analysis also highlights potential observational distinctions between dark energy and modified gravity within this class of theories. A wide range of functional forms for f(Q) gravity have been investigated in the literature. Examples include exponential models Anagnostopoulos et al. 2023 ; Yang et al. 2024 ; Gadbail and Sahoo 2024 ; Su et al. 2025 ; Gadbail et al. 2024 ; Mhamdi et al. 2024 ; Vasquez and Oliveros 2025 ; Ayuso et al. 2025 ; Roy et al. 2025 ; Boiza et al. 2025 , power-law models Beltrán Jiménez et al. 2020a ; Anagnostopoulos et al. 2023 ; Frusciante 2021 ; Bhardwaj et al. 2024 ; Gadbail and Sahoo 2024 ; Sahlu et al. 2025 ; Enkhili et al. 2024 ; Gadbail et al. 2024 ; Kolhatkar et al. 2024 ; Sakr and Schey 2024 ; Wang et al. 2024 ; Mhamdi et al. 2024 ; Yang et al. 2024 ; Ayuso et al. 2025 ; Paliathanasis 2025 ; Roy et al. 2025 ; D’Agostino and Nunes 2022 , hyperbolic tangent forms Anagnostopoulos et al. 2023 ; Su et al. 2025 , logarithmic extensions Nájera et al. 2023 ; Anagnostopoulos et al. 2021 ; Anagnostopoulos et al. 2023 ; Gonçalves et al. 2024 ; Gadbail and Sahoo 2024 ; Sakr and Schey 2024 ; Roy et al. 2025 ; Karmakar et al. 2025 models, and among others Jarv and Pati 2024 ; Bahamonde and Järv 2022 ; Heisenberg 2024 . In addition, several efforts have been devoted to reconstructing the dynamical connection Kar et al. 2022 ; Capozziello and D’Agostino 2022 ; Gadbail et al. 2022 ; Gadbail et al. 2023a ; Naik et al. 2023 ; Mahmood et al. 2024 ; Gadbail et al. 2023b ; Kaczmarek 2024 ; Gadbail et al. 2024 ; Kaczmarek et al. 2025 ; Gadbail et al. 2025 ; El Ouardi et al. 2025 ; Gadbail et al. 2024 ; Yang et al. 2024 ; Yadav et al. 2024 ; Saha et al. 2024 ; Esposito et al. 2022 ; Kang 2021 ; Nojiri and Odintsov 2024 , while other works adopt a designer approach Albuquerque and Frusciante 2022 , and addressing cosmological tensions Sakr and Schey 2024 . In this work, we explore the reconstruction of viable F(Q) cosmological models directly from data, using background probes such as BAO and Pantheon+, along with perturbation-level observables including the effective gravitational constant (\mu) and gravitational-wave propagation modification in amplitude damping (\nu) . While traditional approaches often rely on parameterizations of the dark energy equation of state w(z) , our method instead parameterizes the dark energy density and reconstructs the underlying F(Q) function. This choice is more natural in the context of modified gravity, where the dynamics of the background and perturbations are tied to the fundamental geometric Lagrangian rather than an assumed equation of state of an energy density component. At present, the available data lack the precision necessary to decisively distinguish modified gravity from dark energy at the level of perturbations. In particular, current measurements of growth (f\sigma_{8}) and gravitational wave modifications are not yet sufficient to break the degeneracy between competing scenarios. Nevertheless, our reconstruction delivers model-independent predictions for \mu and \nu , which can serve as benchmarks for upcoming surveys such as DESI, Euclid, SKA, and next-generation gravitational wave detectors LISA, DECIGO. These forecasts highlight the potential of future data to test F(Q) gravity at both the background and perturbation levels, making our method a forward-looking tool to probe the nature of cosmic acceleration. The issue of theoretical stability has also been discussed in detail Hu et al. 2022 ; D’Ambrosio et al. 2023 ; Tomonari and Bahamonde 2024 ; Guzman 2023 ; Gomes et al. 2024 ; Bello-Morales et al. 2024 . In the present work, however, our emphasis lies on the phenomenological implications of f(Q) cosmology, particularly through reconstruction techniques, which are versatile and can be applied across different classes of models. The paper is organized as follows. In Section II , we briefly review the f(Q) gravity and present the background and perturbation equations. In Section III , we describe the datasets and the methodology employed in this work. The results are presented in Section V , followed by a discussion and conclusion in Section VI . II f(Q) gravity We start with the action of f(Q) gravity framework, \mathcal{S}=\int{{\rm d}^{4}}x\sqrt{-g}\Big[-\frac{1}{2\kappa^{2}}f(Q)+\mathcal{L}_{M}(\rm{CDM,b,rad})\Big]\,. (1) with \kappa^{2}=8\pi G_{N} , c=1 and Newton’s constant, G . The matter Lagrangian \mathcal{L}_{M}(\rm{CDM,b,rad}) represents the contributions from relativistic matter (photons, neglecting neutrinos) and non-relativistic matter (cold dark matter and baryons) f(Q) is an arbitrary function of the non-metricity scalar ( Q ). The non metricity scalar is defined as the contraction of the non-metricity tensor Q_{\alpha\mu\nu} , with its conjugate, P^{\alpha\mu\nu} , Q=-Q_{\alpha\mu\nu}P^{\alpha\mu\nu}\,. (2) The non-metricity tensor is given as, Q_{\alpha\mu\nu}=\nabla_{\alpha}g_{\mu\nu} , which by definition incompatible with metric. The non-metricity conjugate is constructed combining its two independent traces Q_{\alpha}=g^{\mu\nu}Q_{\alpha\mu\nu} and \tilde{Q}_{\alpha}=g^{\mu\nu}Q_{\mu\alpha\nu} P^{\alpha}{}_{\mu\nu}=-\frac{1}{2}L^{\alpha}{}_{\mu\nu}+\frac{1}{4}\left(Q^{\alpha}-\tilde{Q}^{\alpha}\right)g_{\mu\nu}-\frac{1}{4}\delta^{\alpha}_{(\mu}Q_{\nu)}\,. (3) The non-metricity conjugate tensor satisfies the variation, P^{\alpha\mu\nu}=-\frac{1}{2}\frac{\partial Q}{\partial Q_{\alpha\mu\nu}}. (4) We assume a homogeneous, isotropic and spatially flat Friedmann-Lemaître-Robertson-Walker Green and Wald 2014 background, whose with metric, {\rm d}s^{2}=-{\rm d}t^{2}+a^{2}(t)\delta_{ij}{\rm d}x^{i}{\rm d}x^{j} . This yields Hubble expansion, H=\dot{a}/a , where a(t) is the scale factor and its derivative are in cosmic time, t . The time evolution of the Hubble parameter satisfies \frac{\ddot{a}}{a}=\dot{H}+H^{2} . Three sets of coordinate choices for the connections has been discussed in the past lituratures. From here onward, we adopt the simplest coordinate choice, Q=6H^{2} Gomes et al. 2024 , that results in the disformation equal to negative Levi-Civita connection for the non in the absence of torsion. The coincident gauge can also be imposed together with the coordinate choice, though not part of this work. II.1 Background Equations Adopting the coordinate choice, Q=6H^{2} , the modified Friedmann equations of f(Q) gravity model in the background are \displaystyle 6H^{2}f_{Q}-\frac{f}{2} \displaystyle= \displaystyle\kappa^{2}\rho_{M}\,, (5) \displaystyle 2\big(12H^{2}f_{QQ}+f_{Q}\big)\dot{H} \displaystyle= \displaystyle-\kappa^{2}(\rho_{M}+p_{M})\,, (6) with f_{Q}\equiv f_{Q}(Q)\equiv\partial f/\partial Q and f_{QQ}\equiv f_{QQ}(Q)\equiv\partial^{2}f/\partial Q^{2} . The total matter density is splitted into pressureless ( p_{m}=0 ) non-relativistic matter density ( rho_{m} ) including cold dark matter and baryons, and relativistic radiation density, \rho_{r} , satisfies \rho_{M}=\rho_{m}+\rho_{r} . p_{m} is the pressure of non-relativistic matter. We are imposing the condition of spatial flatness that is strongly supported by cosmic microwave background radiation observations Ade et al. 2016 . Each species separately satisfies the standard continuity equations, \dot{\rho_{i}}=-3H(\rho+p_{i}) . The Lagrangian with linear Q is equivalent to general relativity Beltrán Jiménez et al. 2018 , that closely follow the background of the universe. Therefore, one can think the non-GR modifications encodes through additional F(Q) functions, satisfies f(Q)=Q+F(Q) . That does not loose the generality, because one could always defne backward F(Q)\equiv f(Q)-Q . In the background equations in form of explicit F(Q) , \displaystyle 3H^{2}+6H^{2}F_{Q}-\frac{F}{2} \displaystyle= \displaystyle\kappa^{2}\rho_{\rm m}+\kappa^{2}\rho_{\rm{r}}\,, (7) \displaystyle 2\dot{H}\big(12H^{2}F_{QQ}+1+F_{Q}\big) \displaystyle+6H^{2}(1+F_{Q})-\frac{6H^{2}+F}{2} \displaystyle= \displaystyle-\kappa^{2}p_{\rm m}-\kappa^{2}p_{\rm r}\,, (8) with the derivatives with respect to Q , F_{Q}\equiv F_{Q}(Q)\equiv\partial F/\partial Q and F_{QQ}\equiv F_{QQ}(Q)\equiv\partial^{2}F/\partial Q^{2} . The above equations can be written in a standard Friedmann-like form, \displaystyle 3H^{2} \displaystyle= \displaystyle\kappa^{2}(\rho_{\rm m}+\rho_{\rm r}+\rho_{\rm de})\,, (9) \displaystyle 2\dot{H}+3H^{2} \displaystyle= \displaystyle-\kappa^{2}(p_{\rm m}+p_{\rm r}+p_{\rm de})\,, (10) where the effective dark energy sector is identified as \displaystyle\rho_{\rm de} \displaystyle\equiv \displaystyle\frac{1}{\kappa^{2}}\left(\frac{F}{2}-6H^{2}F_{Q}\right)\;, (11) \displaystyle p_{\rm de} \displaystyle\equiv \displaystyle\frac{1}{\kappa^{2}}\left(2\dot{H}(12H^{2}F_{QQ}+F_{Q})-\frac{F}{2}+6H^{2}F_{Q}\right)\;. (12) We define the usual critical density of the Universe from From Eq. ( 9 ), \rho_{\rm c}\equiv\frac{3H^{2}}{\kappa^{2}}\;. (13) This allows us to introduce the fractional energy density parameters, \Omega_{\rm m}\equiv\frac{\rho_{\rm m}}{\rho_{\rm c}}\,,\,\Omega_{\rm r}\equiv\frac{\rho_{\rm r}}{\rho_{\rm c}}\,,\quad\Omega_{\rm de}\equiv\frac{\rho_{\rm de}}{\rho_{\rm c}}=\frac{F}{6H^{2}}-2F_{Q}\;. (14) The subscript 0 denotes present-day values of these quantities. Introducing the normalized Hubble parameter, E\equiv H/H_{0} , the modified Friedmann equation Eq. ( 9 ) is recasted as \displaystyle E^{2} \displaystyle= \displaystyle\Omega_{{\rm m}0}(1+z)^{3}+\Omega_{{\rm r}0}(1+z)^{4}+\Omega_{{\rm de}0}R_{\rm de}\, (15) where the dark energy evolution factor is R_{\rm de}\equiv\frac{\rho_{\rm de}}{\rho_{\rm{de}0}}=\frac{F-12H^{2}F_{Q}}{F_{0}-12H_{0}^{2}F_{Q_{0}}}\,.\;, (16) The spatial flatness condition then requires \Omega_{{\rm de}0}=1-\Omega_{{\rm m}0}-\Omega_{{\rm r}0}=\frac{\rho_{\rm de0}}{\rho_{\rm c0}}=\frac{F_{0}}{6H_{0}^{2}}-2F_{Q_{0}}\;. (17) The dark energy equation of state parameter, (w_{\rm de})={p_{\rm de}}/{\rho_{\rm de}} , is defined as \displaystyle w_{\rm de} \displaystyle= \displaystyle\frac{-F+4\left(3H^{2}+\dot{H}\right)F_{Q}+48H^{2}\dot{H}F_{QQ}}{F-12H^{2}F_{Q}}\,.\; (18) Finally, the luminosity distance D_{\rm L}(z) and transverse comoving distance D_{\rm M}(z) can be obtained directly from the dimensionless Hubble expansion rate, E(z)=H(z)/H_{0} , D_{\rm L}(z)=\frac{c}{H_{0}}\int_{0}^{z}\frac{dz^{\prime}}{E(z^{\prime})}\,\equiv(1+z)D_{\rm M}(z)\,., (19) In our work, we numerically solve E(z) from the aforementioned background equations. II.2 Perturbation equations We now summarize the perturbative sector of f(Q) gravity model, focusing on the scalar and tensor modes relevant for large-scale structure formation and gravitational wave propagation. II.2.1 Scalar perturbations The impact of modified gravity on scalar fluctuations can be expressed through the effective gravitational coupling \mu(a,k) . Amendola et al. 2008 , Modification Effective gravitational coupling , \mu , can be written in Fourier space as, In Fourier space, the generalized Poisson equation takes the form Amendola et al. 2008 \displaystyle-\frac{k^{2}}{a^{2}}\Psi=4\pi G_{N}\,\mu(a,k)\rho_{\mathrm{m}}\delta_{\rm m}\,, (20) where \Psi denotes the Newtonian potential, \delta_{\rm m}=\delta\rho_{\rm m}/\rho_{\rm m} is the matter density contrast with respect to the background, and k is the comoving wavenumber. The effective gravitational coupling deviates from standard GR quantifies by the function \mu , which is the function of F(z) . The \Lambda CDM limit is \mu\to 1 . Under the quasi-static approximation, which applies on sub-horizon scales, one obtains Beltrán Jiménez et al. 2020b , \displaystyle\mu(a) \displaystyle= \displaystyle\frac{1}{f_{Q}}\,\equiv\frac{1}{1+F_{Q}}\,. (21) It is worth noting that different conventions for the effective gravitational coupling \kappa^{2}_{\rm eff} exist in the literature, along with corresponding definitions of the critical density \rho_{c} and the effective dark energy density \rho_{\rm de} . A careful discussion of these choices and their interpretation on comparisons with \Lambda CDM is given in Karmakar et al. 2025 . II.2.2 Tensor perturbations Tensor perturbations describe the dynamics of gravitational waves (GWs). In conformal time, the propagation equation for the tensor mode h_{(\lambda)} in f(Q) gravity reads h^{\prime\prime}_{(\lambda)}+2\mathcal{H}\left(1+\frac{{\rm d}\log f_{Q}}{2\mathcal{H}{\rm d}\eta}\right)h^{\prime}_{(\lambda)}+k^{2}h_{(\lambda)}=0. (22) The most general GW propagation equation is given by Nishizawa 2018 h^{\prime\prime}_{ij}+\left(2+\nu\right)\mathcal{H}h^{\prime}_{ij}+(c_{T}^{2}k^{2}+a^{2}\mu^{2})h_{ij}=a^{2}\Gamma\gamma_{ij}. (23) \nu measures deviations from the standard amplitude damping, c_{T} is the GW propagation speed, \mu an effective mass term, and \Gamma\gamma_{ij} possible source contributions. f(Q) models predict c_{T}=1 , i.e. gravitational waves propagate at the speed of light, consistent with current observational constraints. The modification manifests through the damping parameter, \displaystyle\nu \displaystyle= \displaystyle\frac{1}{\mathcal{H}}\frac{{\rm d}\log f_{Q}}{{\rm d}\eta}\equiv\frac{1}{H}\frac{{\rm d}\log f_{Q}}{{\rm d}t}\,, (24) \displaystyle= \displaystyle 12\dot{H}\frac{f_{QQ}}{f_{Q}}\;, \displaystyle= \displaystyle 12\dot{H}\frac{F_{QQ}}{1+F_{Q}}\;. (25) The prime denotes differentiation with respect to conformal time \eta , follows relation dt=a,d\eta . III Modeling In the present work, in contrast to the usual approach, where in f(Q) is assumed to be a specific function of Q , we explore the ‘model-independent’ approach, where we do not assume any specific form of f(Q) , but rather parametrize the dark energy contribution. To this end, we start by modeling the dark energy density ( R_{\rm de} ) in eq. 9 as a 3^{\rm rd} order truncated Taylor expansion 1 1 1 We assume that the 3^{\rm rd} order Taylor expansion has sufficient radius of convergence to ‘correctly’ estimate the dark energy density in the redshift range ( a\leq 0.33 ) of data available, which is validated also by our estimation of the DE EoS which is in agreement with the results presented in Abdul Karim et al. 2025 . in the scale factor a as \frac{\rho_{\rm de}}{\rho_{\rm de0}}=1+\alpha_{1}(a-1)+\alpha_{2}(a-1)^{2}+\alpha_{3}(a-1)^{3}\,, (26) where \rho_{\rm de0} is the present value of the dark energy density, and \alpha_{i} are the model parameters. Then \Omega_{\rm de0}=\rho_{de0}/\rho_{c0} , is straight away inferred from the closure condition, \Omega_{\rm de0}=1-\Omega_{\rm m0}-\Omega_{\rm r0} , where \Omega_{\rm m,r} are the present day matter and radiation densities, respectively. This parametrization enusre that the dark energy density is a smootly varying function with no sharp discontinuities. Also it is not restricted to be positive definite at all times. Once the dark energy density is obtained through the Bayesian inference, we then translate the dark energy density to the F(Q) function, solving the first order ordinary differential equation following Eq. ( 11 ), F(Q)=6H^{2}\left(\Omega_{\rm de}+2F_{Q}(Q)\right)\,, (27) where F_{Q} is the derivative of the function F(Q) . We do so for each of the sample in the posterior distribution, solving the above differential equation numerically, assuming Q=6H^{2} as coordinate choice.In order to solve the above first order differential equation one can make assumptions regarding two initial conditions, either on the value of F(Q) itself or it’s derivative F_{Q}(Q) , which imply a possibility for two different approaches, where in we can assume one to obtain the other. The two approaches then are: • Approach-I: F(Q)|_{z\gg 0}\to 0 , implies the early-time is described by a GR like scenario with no cosmological constant ( \Lambda\to 0 ). Essentially a Einstein-de Sitter universe that evolves into an late-time accelerating phase, following the redsfhit evolution of F(Q(z)) . • Approach-II: F_{Q}(Q)|_{z\gg 0}\to 0 , implies the early-time is described by a GR like scenario with non-zero cosmetological constant ( \Lambda\to const) and a small derivative, allowing later redsfhit evolution. Here the cosmological constant can remain a constant or evolve. We remain with the approach-II as in a cosmological constant is present which can however also imply \Lambda=0 , within in the posterior limits. We also test the Approach-I that tends to prouduce similar resutls however with sliglhtly larger deviations from \Lambda CDM, unless the inital redshift is choosen to be sufficiently large. In the analysis presented here, we assume the initial condition at z=5(a\sim 0.2) , which is well outside the range of observational data, alos being conservative for the radius of convergence of the taylor expansion to not fail. IV Data and Analysis We briefly describe the observational datasets and the statistical analysis framework, which is very close as implemnted in Karmakar et al. 2025 ; Tyagi et al. 2025 and elaborated thenin. We utlise a combination of Type Ia supernovae (SNe), baryon acoustic oscillations (BAO), and early-universe priors from the CMB. • Supernovae (SNe): We use the Pantheon+ sample of 1550 SNe Ia ( 0.001\leq z\leq 2.26 ) Scolnic et al. 2022 and the DESyr5 sample of 1830 SNe Ia ( 0.02\leq z\leq 1.13 ) Möller et al. 2024 . Both are treated as uncalibrated in the joint analysis: Pan+ fits the absolute magnitude M_{b} as a free parameter, while DESyr5 uses analytic marginalization. The likelihood is Gaussian in the distance modulus residuals, \mathcal{L}_{\rm SNe}\propto\exp\!\left[-\tfrac{1}{2}\sum_{i}\frac{(\mu_{\rm th}(z_{i})-\mu_{\rm obs}(z_{i}))^{2}}{\sigma_{\mu,i}^{2}}\right], where \mu_{\rm th}(z)=5\log_{10}[D_{L}(z)/\text{Mpc}]+25 and D_{L}(z)=(1+z)D_{M}(z) . • Baryon Acoustic Oscillations (BAO): We use DESI DR2 measurements of D_{M}(z)/r_{d} and D_{H}(z)/r_{d} in six bins spanning 0.1\leq z\leq 4.2 Abdul Karim et al. 2025 . The BAO likelihood is \mathcal{L}_{\rm BAO}\propto\exp\!\left[-\tfrac{1}{2}\sum_{i}\Delta\vec{\mu}_{i}^{\,T}C_{i}^{-1}\Delta\vec{\mu}_{i}\right], where \Delta\vec{\mu}_{i}=(D_{M,i}^{\rm obs}-D_{M,i}^{\rm th},\,D_{H,i}^{\rm obs}-D_{H,i}^{\rm th}) . • Inverse Distance Ladder (IDL) Priors: To anchor BAO we adopt early-universe priors on the sound horizon r_{d} and recombination Hubble parameter H_{\rm rec} from Planck 2018, with covariance from Haridasu et al. 2021 (equivalent to the approach in Lemos and Lewis 2023 ; Adame et al. 2025 ; Lodha et al. 2025 ; Lynch and Knox 2025 ). The likelihood is Gaussian, \chi^{2}_{\rm IDL}=\Delta\bm{\mu}_{\rm IDL}^{T}\,C^{-1}\,\Delta\bm{\mu}_{\rm IDL}, with \Delta\bm{\mu}_{\rm IDL}=\{\Delta H_{\rm rec},\Delta r_{d}\} . These datasets are complementary in redshift coverage and precision, and together provide a robust probe of the expansion history. The joint likelihood is \ln\mathcal{L}=\ln\mathcal{L}_{\rm SNe}+\ln\mathcal{L}_{\rm BAO}+\ln\mathcal{L}_{\rm IDL}. (28) The posterior sampling is performed with Nautilus Lange 2023 , and ingerence of constraints is performed using GetDist Lewis 2025 . Figure 1: Posterior distributions of the dark energy density ( top ) as function modeled as a 3rd order Taylor expansion in the scale factor a as described in Eq. ( 26 ) and the corresponding DE equation of state ( bottom ) parameter w_{\rm de} , defined in Eq. ( 18 ), as a function of the redsfhit. The dashed line corresponds to the cosmological constant case, w_{\rm de}=-1 . The left and right panels correspond to the dataset combination DESI+Pan + and DESI+DESyr5, respectively. V Results and Discussion We begin by firstly presenting the posterior distributions of the dark energy density from the joint analysis of SNe and BAO datasets. In Figure 1 , we show the posterior distribution of the reconstructed dark energy density (top panel) and the corresponding equation of state parameter w_{\rm de} (bottom panel) as a function of the redshift. The left and right panels correspond to the dataset combination DESI+Pan + and DESI+DESyr5, respectively. The dashed line in the top panel corresponds to the cosmological constant case, \rho_{\rm de}=\rho_{\rm de0} , which accordingly in the bottom panel corresponds to w_{\rm de}=-1 . The shaded region are the 1\sigma and 2\sigma C.L. of the reconstructed dark energy density and equation of state parameter w_{\rm de} . We note that there is a significant difference between the dark energy density reconstructed from the two dataset combinations driven by the different SNe datasets as expected Adame et al. 2025 ; Abdul Karim et al. 2025 ; Brodzeller et al. 2025 . The Pan + dataset shows a mild deviation from the cosmological constant case, \rho_{\rm de}=\rho_{\rm de0} , at lower redshift, z\lesssim 1.5 , while the DESyr5 dataset shows a larger deviation as is also reported in Möller et al. 2024 . This difference is also reflected in the posterior distribution of the equation of state parameter w_{\rm de} , where the DESyr5 dataset shows a larger deviation from the cosmological constant case, w_{\rm de}=-1 , at lower redshift, z\lesssim 1.5 . Also corroborating the detection of the phantom crossing, which has received a overwhelming attention. This difference could possibly be attributed to the different systematics in the two datasets, which are yet to be strictly assessed in the relatively newer DESyr5 compilation Efstathiou 2025 . As can be seen in fig. 1 , the posterior distribution of the reconstructed dark energy density is very well constrained with deviations of the order of 1-2\sigma from the cosmological constant case, w_{\rm de}=-1 , when using the DESI and Pan + dataset. Note that when parameterizing the DE EoS instead of the DE density as an higher-order expansion is expected to reduce the significance of the detection. This is very well understood and expected in the higher order expansions of the EoS, see for instance Nesseris et al. 2025 at higher redsfhit and shows reduced evidence for a phantom crossing as expected from the standard CPL parametrization, as reported in Brodzeller et al. 2025 ; Elbers et al. 2025 . However, in our analysis, we parametrize the DE density, which is a more direct constituent of the expansion rate, and the EoS itself is a later translation of background quantities. And as expected we find a strong evidence for a phantom crossing, w_{\rm de}<-1 , at lower redshift, z\lesssim 1.5 , in agreement with the results presented in Abdul Karim et al. 2025 ; Brodzeller et al. 2025 , when utilizing the DESI+DESyr5 dataset combination. V.1 Reconstruction of the f(Q) function and modified gravity parameters Having estimated the ’model-independent’ evolution dark energy, we now turn to the reconstruction of the f(Q) function from the dark energy density, following the approach described in Section III . The posterior distribution of the F(Q) as function of redsfhit is shown in fig. 2 . We find each of these quantities well-constrained essentially implying a very good correspondence between the reconstructed dark energy density and the unassumed f(Q) function. The reconstructed F(z) function is consistent with the cosmological constant case, F(z)=6H^{2}\equiv Q , within 1\sigma level all redshifts. While at the higher-redshifts the posterior distribution of F(z) is also very well consistent with the case, F(z)=0 ( F_{0} ), at \sim 1\sigma level, for both the dataset combinations DESI+Pan + and DESI+DESyr5. However, at lower redshift, z\lesssim 1.5 , the posterior distribution of F(z) shows a strong deviation from this F(z)=0 case, indicating the presence of at least a cosmological constant like term, ensuring late-time acceleration. The deviation is more pronounced for the DESI+DESyr5 dataset combination, which is also reflected in the reconstructed dark energy density and equation of state parameter w_{\rm de} as shown in fig. 1 . We find a constraint of F_{0}/10^{3}=5.95^{+1.15}_{-1.13} and asymptotic high-redshift value of F_{z\gg 0}/10^{3}=5.20^{+4.05}_{-3.83} , for the dataset combination DESI+DESyr5. Similarly for the DESI+Pan + combination we find F_{0}/10^{3}=5.64^{+1.15}_{-1.08} and F_{z\gg 0}/10^{3}=2.73^{+3.95}_{-3.75} , showing a lower value of the latter wrt the DESI+DESyr5 dataset. Scalar perturbations: Then the reconstruction of running of the gravitational constant \mu=1/(1+F_{Q}) 2 2 2 The posterior distribution of F_{Q} function is mostly consistent with the, F_{Q}\sim 0 , within 1\sigma level at higher redshift and the lower z\lesssim 1.5 . The bottom panel of fig. 2 shows the full F(z) which deviates significantly at lower redshift, z\lesssim 1.5 indicating an acceleration, while slowly tending to F(z>3)\to 0 at higher redshifts. can be inferred from the reconstructed F_{Q} function. The posterior of the effective gravitational coupling \mu is shown in the center-panel of fig. 2 . The posterior distribution of \mu is consistent with the cosmological constant case, \mu=1 , well within 1\sigma level at higher redshift and shows a deviation at lower redshift, z\lesssim 1.5 . We find a 1\sigma deviation locally, z=0 , indicating a deviation from the standard \Lambda CDM model. Our reconstruction indicates a mild increase of the effective gravitational coupling reaching \mu\sim 1.01\pm 0.02 which is a 2\% increase wrt the Newtonian G_{\rm N} . Interestingly, this inference is completely consistent between the two datasets combinations DESI+Pan + and DESI+DESyr5. This finding also implies that the \sim 2\% variation in the gravitational constant can yield very little advantage in addressing the H_{0} and S_{8} -tensions (see Di Valentino et al. 2025 for a review). Our results can be contrasted against the recent findings in Boiza et al. 2025 , where the implications of f(Q) gravity for the tensions was discussed assuming specific parametric model of F(Q) using the similar dataset combinations. Tensor perturbations: Finally, we reconstruct the damping parameter \nu from the reconstructed F_{Q} and F_{QQ} functions using Eq. ( 25 ). The reconstructed running parameter \nu is shown in the right-panel of fig. 2 . The posterior distribution of \nu is mostly consistent with the cosmological constant case, \nu=0 , within 1\sigma level within z\lesssim 2.5 where data is available for the DESI+Pan + combination. However, we find the median curve of \nu(z) to be negative, \nu\sim-0.02 at z\sim 1 reminiscent of the predictions made in a exponential model fro F(Q ) or the inverse logarithmic model we introduced in Karmakar et al. 2025 , which then transitions to positive values at z\sim 0.2 . We find a constraint of \nu=0\pm 0.06 after the said oscillations in redshift. Figure 2: Top row corresponds to the data combination DESI+Pan + and the bottom corresponds to the DESI+DESyr5. We show the posterior distribution of F(z) ( left ), the effective gravitational coupling \mu ( center ) and the damping factor \nu ( right ) as a function of the redshift. The dashed line in the latter two corresponds to the cosmological constant case, \mu=1,\,\nu=0 . The shaded region corresponds to the 1\sigma and 2\sigma C.L. of the reconstructed quantities. Now turning to the most crucial finding of the current analysis, the reconstructed \nu using the DESI+DESyr5 dataset shows a strong signature for a deviation from GR like scenario. This signature, interestingly remains similar in nature to that from DESI+Pan + , however, more pronounced. As shown in the right-bottom panel of the fig. 2 , we find a steep increase in the redshift range 0\lesssim z\lesssim 0.5 reaching from \nu\sim-0.05 at z\sim 0.5 to \nu\sim 0.15 at z=0 . This indicates a strong deviation from the GR like scenario, \nu=0 , at >2\sigma level locally, z\sim 0 . This finding is also consistent with the strong deviation from the cosmological constant case, w_{\rm de}=-1 , as shown in fig. 1 . We find a constraint of \nu=0.18^{+0.08}_{-0.07} at 68\% C.L., which is a \gtrsim 2\sigma deviation from the GR scenario, \nu=0 . We stress once again, that in our current analysis, we find that the DESI+Pan + dataset combination does not show such a strong deviation from GR like scenario, \nu=0 , which could be attributed to the different systematics in the two SNe datasets as discussed earlier. Note that most traditional parametric models do no allow for a transition of the \nu which remain either positive or negative throughout the cosmic history. At a face-valve, one can immediately infer that that the nature of the f(Q) function is such that it allows for a transition of the running parameter \nu at lower redshift and hence demonstrates the inadequacy of several proposed models and subsequent tests performed in literature. On the other hand, it remains an important prediction for the gravitational wave propagation that can be crucial to distinguish between different modified gravity models and more importantly form dark energy models within GR. In the context of phenomenological hints/evidence for dynamical dark energy, such a behavior sourced by fundamental modified gravity model will display varied predictions in the scalar and tensor sectors. The current analysis indicates that the DESI+DESyr5 dataset combination shows a strong evidence for a dynamical dark energy, which is also reflected in the reconstructed f(Q) function and the modified gravity parameters \mu and \nu . VI Conclusions In the work, we have performed a ’model-independent’ reconstruction of the f(Q) cosmology, as an example, from the latest SNe and BAO data. The analysis has been performed to assess the implications for the modified gravity theories in the context of the recently reported hints for a dynamical dark energy. We have used two different SNe datasets, the Pan + and the DESyr5, in combination with the latest BAO measurements from DESI-DR2. We find that the reconstructed F(Q) function is consistent with the cosmological constant case, f(Q)=Q+\Lambda , in agreement with \Lambda=0 within 1\sigma level at high-z. While showing a significant evidence at lower redshifts, z\lesssim 1.5 being different from F_{0}\equiv\Lambda=0 , necessary to source the late-time acceleration. We then estimated the deviation of effective gravitational coupling \mu and the damping parameter of GW propagation \nu from the reconstructed f(Q) function. We find that \mu is consistent with \mu=1 , within 1\sigma level, while showing a mild increase at lower redshifts, reaching \mu\sim 1.02\pm 0.02 at present time. The gravitational wave amplitude damping parameter \nu is mostly consistent with standard scenario, \nu=0 , within 1\sigma level within z\lesssim 2.5 , but shows a negative median value at higher redshifts, reminiscent of predictions made in inverse logarithmic model of f(Q) using the DESI+Pan + dataset Karmakar et al. 2025 . However, the DESI+DESyr5 dataset combination shows a strong deviation from the GR like scenario, \nu=0 , at >2\sigma level locally, z=0 , reaching \nu=0.18^{+0.08}_{-0.07} at 68\% C.L. This finding is also consistent with the strong deviation from the cosmological constant case, w_{\rm de}=-1 , as shown in fig. 1 . This in fact implies a possibility for future gravitational wave standard siren measurements to further constrain the \nu parameter and hence the f(Q) function, which can distinguish between different modified gravity models and dark energy models within GR, proving equivalent dynamical dark energy behavior. Our results on the one hand provides necessary insights into the nature of the f(Q) function and its implications for modified gravity theories in the context of model building going beyond the simple parametric forms explored so far in literature. On the other hand, we find that the results are dataset dependent, also highlights the importance of understanding and mitigating systematics in SNe datasets, as different compilations can lead to varying conclusions. We have explored a framework, that can be extended to other modified gravity models to test the phenomenological implications of the hints for a dynamical dark energy, through precise perturbation and gravitational wave data to arrive in the near future. Surveys with improved precision and control over systematics will be crucial to further elucidate the nature of dark energy and its connection to modified gravity theories.