Forecasting the Area of the Bakhtegan and Tashk Lake Using Remote Sensing and Climatic Factors

Recurring droughts at various intensity leads to decline in streamflow and wetland area. In this study, using Landsat satellite images, temporal changes in the area of the Bakhtegan and Tashk lake from 1986 to 2019 were investigated. Also, based on the Standard Precipitation Index (SPI), the effect of precipitation falling on the lake area as well as the total inflow to the lake were studied. Then, by generating land-use maps, land-use changes in Bakhtegan and Tashk basin area were analyzed. In addition, a model for forecasting lake area (of next year) on the basis of lake area, total inflow, and annual precipitation falling on the lake basin area associated with the previous year was presented. The results showed that in the study period, the minimum lake area occurred in 2011 and was equal to 17.038 square kilometers, which shows a decrease of 98% compared to the maximum area corresponding to the year 1993. Furthermore, 12-month SPI in Bakhtegan and Tashk Lake basin indicated that the meteorological drought index directly affects the changes of the lake area. However, from 2006 to 2012, the inflow to the lake has severely decreased due to the increasing human interventions within the basin such that, during this period, lake surface area does not strongly respond to SPI variations. The results of the forecast model showed that the coefficient of determination between neural network model output (lake area) with observed lake area corresponding to satellite images was equal to 0.72. As a result, the forecast model could acceptably predict the lake area.