Simulation of peri-urban growth dynamics using weights of evidence approach

Forecasting Spatio-Temporal Dynamics on the Land Surface Using Earth Observation Data—A Review

“…Section 3.3.3). Since MC modeling as such is not spatially explicit [40,41], these methods are in most cases integrated with CA models that also consider the state of the neighboring pixels when calculating the transition probabilities and allocate state changes according to a local suitability map [40][41][42][43][44][56][57][58][66][67][68][69][70][71]158]. To improve the accuracy of these models, the transition probabilities and suitability (i.e., probability whether and where LULC change will occur) are frequently expressed as functions of multiple explanatory variables such as topography, socio-economic metrics and distance functions.…”

“…The most important applications in EO-based forecasting of the anthroposphere are LULC (54%) and crop yield (40%). In most LULC simulations, forecasts focus on urban sprawl or LULC change in an urban environment, simulating more general LULC maps of urban centers [40][41][42][43][44][45][46][47][48][49][50][51][52][53][54][55] or binary urban/non-urban masks [56][57][58][59][60][61][62][63][64]. Musa et al [36] reviewed urban modeling studies and showed that modeling approaches based on CA are most popular in the scientific literature due to their flexibility and ability for spatially explicit simulation.…”

“…Self-learning iterative and time series methods are used more frequently. While time series-based methods and the hybrid NARX model are used especially for shorter forecasting horizons up to six years [160,161,179], SI are applied for longer lead times up to ten years [52,56,65,66,69,74,75,80,82,83,132,133,157]. Even longer lead times are achieved especially with SI and RP methods.…”