Snow cover variations - Aysén river - 2024
The application provides an interactive visualisation of monthly variations in snow cover during a defined year. By clicking on the appropriate column of the month in the chart, the corresponding monthly raster is shown in the map window. Different base maps can be added to the map composition.
Computational Network Model COMUNIDAD · WP4 · D4.4 · Aysén River Basin, Chile
Forage Biomass (x₁)
N(t)-modulatedLivestock Density (x₂)
NormalisedEmployment (x₃)
NormalisedNet Income (x₄)
NormalisedPhase Portrait — All State Variables
t = 0 → 100 days
Biomass x₁
Livestock x₂
Employment x₃
Income x₄
Multi-Scale Hierarchical Architecture
Two-Layer Complex Network · Aysén BasinMICRO
Pixel / Agent Scale — EO Data Ingestion
High-resolution Sentinel-2 imagery → Snow Persistence N(t) via NDSI + Random Forest binary classification.
Multi-sensor LULC (Sentinel-1 SAR + Sentinel-2 optical + DEM topographic) → Patch land-use classification.
Ingested via OGC WMS/WCS → PostgreSQL/PostGIS spatial database.
⇣ N(t), Lᵢ(t) forcing ⇣
MESO
Patch Scale — Functional Subsystems
Territory discretised into 9 patch types: Agricultural (A), Livestock (G), Forestry (F),
Tourism-General (T), Tourism-Fishing (P), Rural (R), Urban (U), Salmon (S), Unused (Z).
Each patch governed by 4-variable ODE system: xᵢ = [biomass, livestock, employment, income].
Snow forcing N(t) modulates logistic growth. Inter-patch coupling via mobility matrix wᵢⱼ.
⇣ Aggregated R(t) ⇣
MACRO
Basin Scale — Aggregated Dynamics
State vector: [L, N, F, H] = [Population, Non-renewable resources, Fisheries, Social index].
Renewable resources R = Σ(MESO patches) → bottom-up coupling.
Captures slow structural constraints: demographic trends, resource depletion, social feedback loops.
Layer 1: Uncontrolled SLW System (Stochastic)
Snow-Land-Water environmental forcing. Meteorological drivers + Copernicus Sentinel-2 time series →
continuous dimensionless forcing function N(t) that modulates vegetation growth rates and hydrological availability.
Layer 2: Controlled Land-Use System (Deterministic + SDE)
Socio-economic dynamics discretised into functional patches. Each node governed by ODEs with
multiplicative noise terms (SDE extension) for climate uncertainty. Network edges represent mobility
and economic flow matrices between nodes.
Source: Deliverable D4.4 — Annexe 1 & 2 · MatCont/MATLAB validation · COMUNIDAD Project (GA 101136910)
ODE System — Mathematical Formulation
Dimensionless form · MatCont validatedMESO Level — Agricultural-Livestock Patch (Ñirehuao)
dx₁/dt = (1 + α·N)·x₁·(1 − x₁) − β·x₁·x₂ − (δ₁ + δₜ)·x₁
dx₂/dt = γ·x₁·x₂ − δ₂·x₂ + π
dx₃/dt = η·x₂·(1 − x₃) − δ₃·x₃
dx₄/dt = ρ·x₂ − δ₄·x₄ − χ
Where: x₁ = normalised forage biomass, x₂ = livestock density, x₃ = employment rate, x₄ = net income.
N(t) = snow persistence forcing from Sentinel-2. Parameters calibrated for Ñirehuao sector, Aysén.
MACRO Level — Basin Scale
dL/dt = γ·(P − σ)·L
dN/dt = −α_j·β_n·L
dF/dt = ρ·F·(1 − F/K_f) − α·β_f·L·F
dH/dt = q·G_e·(m₁R + m₂N)·H·(1−H)·(1−L/Lᵢ)·(L/Lₐ−1)·L
Where: L = total population, N = non-renewable resources, F = fisheries, H = social index.
R = renewable resources (aggregated from MESO). Basin-wide slow dynamics providing structural constraints.
Network Coupling Structure
dxᵢ/dt = fᵢ(xᵢ, pᵢ, ηᵢ(t)) + Σⱼ wᵢⱼ·(xⱼ − xᵢ)
General inter-patch coupling: wᵢⱼ represents mobility and economic flow weights between patches.
The directed network ensures information flows across scales: MICRO → MESO → MACRO.
COMUNIDAD Project · Horizon Europe · GA 101136910 · Connecting Europe and Latin America