Weighted Point Configurations with Hyperuniformity: An Ecological Example and Models
Ayana Ezoe, Makoto Katori, Tomoyuki Shirai · 2025
Read Weighted Point Configurations with Hyperuniformity: An Ecological Example and Models by Ayana Ezoe, Makoto Katori, Tomoyuki Shirai (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-12807.txt.
Introduction
Ayana Ezoe, Makoto Katori, and Tomoyuki Shirai study desert bush patterns as an example of hyperuniformity, a state where density fluctuations are unusually suppressed at large scales. In dry lands, competition for water may push plants into such regular spacing. They digitized bush positions in the field, weighted by bush size.
Ignoring size, the pattern was not hyperuniform, but counting the masses made it so. That distinction between plain and weighted patterns is the heart of the paper. They propose models where repeated random thinning of points and merging of masses turn a plain random pattern into a hyperuniform weighted one. Simulations and data together stress that space and mass must be strongly linked for the effect.
Catalog record
| CREATOR | Ayana Ezoe, Makoto Katori, Tomoyuki Shirai |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 52 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-12807
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Weighted Point Configurations with Hyperuniformity: An Ecological Example and Models Ayana Ezoe, Makoto Katori, Tomoyuki Shirai Abstract Random point configurations are said to be in hyperuniform states, if density fluctuations are anomalously suppressed in large-scale. Typical examples are found in Coulomb gas systems in two dimensions especially called log-gases in random matrix theory, in which points are repulsively correlated by long-range potentials. In infertile lands like deserts continuous survival competitions for water and nutrition will cause long-ranged repulsive interactions among plants. We have prepared digital data of spatial configurations of center-of-masses for bushes weighted by bush sizes which we call masses. Data analysis shows that such ecological point configurations do not show hyperuniformity as unmarked point processes, but are in hyperuniform states as marked point processes in which mass distributions are taken into account. We propose the non-equilibrium statistical-mechanics models to generate marked point processes having hyperuniformity, in which iterations of random thinning of points and coalescing of masses transform initial uncorrelated point processes into non-trivial point processes with hyperuniformity. Combination of data analysis and computer simulations shows the importance of strong correlations in probability law between spatial point configurations and mass distributions of individual points to realize hyperuniform marked point processes. 1 Introduction Structures and distribution of ordered and disordered configurations of points in a space have been important research subjects in physics, where the points represent the locations of atoms or molecules in gases, liquids, glasses, quasicrystals, and crystals. If we consider a classical ideal-gas, the spatial configurations of molecules are completely random and the statistical ensembles of such uncorrelated point configurations is identified with the Poisson point process (PPP) studied in mathematics [ 1 , 2 ] . In mathematical physics and probability theory spatial statistics of (random) configurations of points are simply called a (random) point process in general [ 1 , 2 ] . Correlated point processes have been also extensively studied in statistical mechanics for Coulomb gas systems, which are especially called the log-gases in one- and two-dimensional spaces [ 3 ] . It is known that they are realized as the eigenvalue distributions of Hermitian and non-Hermitian random matrices [ 3 , 4 , 5 , 6 , 7 ] and as the zeros of random analytic functions [ 8 , 9 ] . Recently in condensed matter physics and related material sciences, correlated particle systems are said to be in a hyperuniform state , when density fluctuations are anomalously suppressed in large-scale [ 10 ] . Consider a bounded domain with linear size \ell in the d -dimensional space, which we call an observation window . In this window we measure the mean number of points \langle N_{\ell}\rangle and its variance \sigma^{2}_{\ell} . In a hyperuniform system, as the window size \ell\to\infty , \sigma^{2}_{\ell} grows more slowly than \langle N_{\ell}\rangle which is proportional to the window volume. Typical disordered systems (e.g., liquids and structured glasses) have the standard scaling behavior \sigma^{2}_{\ell}\simeq\ell^{d} as \ell\to\infty , and hence they are not hyperuniform. Periodic point configuration (e.g., lattice points of perfect crystals) are obviously hyperuniform systems, since the number fluctuations are concentrated near the window boundary and hence have the surface-area scaling \sigma^{2}_{\ell}\simeq\ell^{d-1} as \ell\to\infty . We are interested in the hyperuniform point processes, which are not perfect lattice, but are randomly distributed following some probability laws. A variety of examples of disordered or random point processes showing hyperuniformity have been reported in the review paper by Torquato [ 10 ] . A typical example on the two-dimensional plane, which has been well studied in mathematical physics, is the complex Ginibre point process (GPP) [ 4 ] . This is realized as the statistical ensemble of eigenvalues of non-Hermitian random matrices on the complex plane \mathbb{C} , which can be identified with the two-dimensional space \mathbb{R}^{2} . First we consider an n\times n matrix whose entries are all independently and identically distributed (i.i.d.) complex standard Gaussian random variables. We have a nearly uniform distribution of eigenvalues in a disk centered at the origin with radius proportional to \sqrt{n} on \mathbb{C} . If we consider the n\to\infty limit, we have a translation-invariant point process on \mathbb{C} [ 4 ] . It was proved that the obtained GPP is hyperuniform with \sigma^{2}_{\ell}\simeq\ell [ 11 ] . The origin of such hyperuniformity in a random point configuration is long-ranged repulsive interaction between any pair of points. As a matter of fact, GPP is a determinantal point process (DPP) [ 8 ] , for which we can show that a logarithmic repulsive potential acts between any pair of points [ 3 , 12 ] . It should be remarked that DPP is also called fermion point process [ 13 ] , since the determinantal structure of correlation functions is essentially equivalent with the Slater determinant describing wave function of fermionic particle systems. We notice that GPP has been studied as a typical example of strongly correlated point process in two dimensions in non-Hermitian physics [ 14 ] , spectral theory [ 15 ] , random matrix theory [ 16 ] , and so on, and used in many applications (see, for instance, the references [ 17 , 18 , 19 , 20 ] ). Figure 1 : A sample of point configuration of center-of-masses of bushes in a desert. The sizes of area are indicated not by real lengths in meters but by numbers of pixels for the digital data; L_{x}=1626 , L_{y}=895 . The total number of points is N=9853 and the density is \rho\fallingdotseq 6.77\times 10^{-3} . The bush-size (mass) distribution is represented by colors of each points; the points with masses 1\leq M(\bm{X}_{j})\leq 49 are green, the points with 50\leq M(\bm{X}_{j})\leq 99 are blue, and the points with M(\bm{X}_{j})\geq 100 are red. See Section 3 showing how to produce such digital data. In the present paper, we study the spatial distributions of bushes in deserts. In infertile lands plants can survive only if they have continued to beat the competitions to obtain sufficient amount of water and nutrition. It is expected that such survival competitions will cause strong and long-ranged repulsive interactions among plants. Figure 1 shows one sample of point process obtained from the real data of bushes distributed in the 2195.1 m \times 1208.25 m area \Lambda in a desert found in the Talampaya Natural Park in Argentina. There we plotted N=9853 points \bm{X}_{j}\in\Lambda , j=1,2,\dots,N , each of which gives the center-of-mass of a bush. As explained in detail in Section 3 , we will define a size for each bush, which we call a mass and write as M(\bm{X}_{j}) , j=1,2,\dots,N in the present paper. We have obtained several real samples of point configurations of center-of-masses of bushes \mathcal{P}=\{\bm{X}_{j}\}_{j=1}^{N} weighted by masses \{M(\bm{X}_{j}):\bm{X}_{j}\in\mathcal{P} } in deserts. Such weighted point configurations are called marked point processes in mathematical literature [ 1 , 2 ] . The partial information of masses is shown by coloring of points in Fig. 1 . We can see in Fig. 1 that the point configuration tends to sparse around the points whose masses M(\bm{X}_{j})\geq 100 (the bushes with large sizes indicated by red points). On the other hand, the points with small masses M(\bm{X}_{j})\leq 49 (the bushes with