Vulnerability of Land Conversion to Pond Aquaculture
The following layers were developed by relating historical transitions to pond aquaculture to covariates such as elevation, proximity to the coast, proximity to existing pond aquaculture and situational information (such as land cover and soil type) as encapsulated in Landsat TM reflectance data from 1988-1990 for Thailand, Cambodia and Vietnam. Historical transitions were established by comparing the 2013-2015 land cover to a classification of Landsat TM data from 1988‐1990 for mangroves, cropland, stabilized sand and sea salt production – the primary classes from which pond aquaculture is developed. The empirical modeling procedure used was the Multi-Layer Perceptron (MLP) Neural Network in the TerrSet software system (Clark Labs). The resulting maps express the vulnerability to transition from each of these four classes to pond aquaculture on a 0-1 scale. Technically, the numbers represent the probability of transition assuming an equal prior probability of transition or persistence. A simpler understanding is that it represents a measure of the vulnerability to transition irrespective of how prevalent the rate of transition is.
Risk of Land Conversion to Pond Aquaculture by 2050
These layers express the probability of conversion to pond aquaculture by 2050 for Thailand, Cambodia and Vietnam. They are derived from the vulnerability maps (above) by means of prior correction that recognizes the expected quantity of transition given the historical rate of change. The procedure is based on the basic Markovian assumption of a constant rate of change and the rate of change determined from 1988-1990 to 2013-2015 (a 25 year span).