Google Earth Engine for improved spatial planning in agricultural and forested lands: A method for projecting future ecological quality

Abstract

This research aims to use Google Earth Engine and the CA algorithm in open-source software to generate multiple scenarios for predicting future ecological quality and evaluating the spatial planning of agricultural and forested lands. The goal is to determine the most accurate scenario and use it to assess a spatial planning map of these lands. Three different Remote Sensing-based Ecological Index (RSEI) scenarios were simulated using Landsat images and MOLUSCE (Modules for Land Use Change Evaluation) was used to predict and select the most accurate scenario. The results showed that a future RSEI projection based on data from a 7-year interval (2007 and 2014) had the highest accuracy, as indicated by 83.79 percent correctness and an overall Kappa value of 0.72, compared to the other two alternative projections (based on data from a 4-year interval: 2013 and 2017; and a 12-year interval: 1997 and 2009). The selected scenario was then used to project an RSEI map for 2028, which is expected to show an expansion of areas with poor ecological quality, particularly near urban centers. The “poor” and “very poor” areas of the projected RSEI map were used to assess a spatial planning map of agricultural and forested lands for 2031, resulting in the identification of planned agricultural and forested areas at high risk for poor ecological quality in the future. Local governments and urban planners are advised to prioritize the conservation of these areas.