Forecasting GRACE Data over the African Watersheds Using Artificial Neural Networks

Abstract

The GRACE-derived terrestrial water storage (TWS GRACE ) provides measurements of the mass exchange and transport between continents, oceans, and ice sheets. In this study, a statistical approach was used to forecast TWS GRACE data using 10 major African watersheds as test sites. The forecasted TWS GRACE was then used to predict drought events in the examined African watersheds. Using a nonlinear autoregressive with exogenous input (NARX) model, relationships were derived between TWS GRACE data and the controlling and/or related variables (rainfall, temperature, evapotranspiration, and Normalized Difference Vegetation Index). The performance of the model was found to be "very good" (Nash-Sutcliffe (NSE) > 0.75; scaled root mean square error (R*) < 0.5) for 60% of the investigated watersheds, "good" (NSE > 0.65; R* < 0.6) for 10%, and "satisfactory" (NSE > 0.50; R* < 0.7) for the remaining 30% of the watersheds. During the forecasted period, no drought events were predicted over the Niger basin, the termination of the latest (March-October 2015) drought event was observed over the Zambezi basin, and the onset of a drought event (January-March 2016) over the Lake Chad basin was correctly predicted. Adopted methodologies generate continuous and uninterrupted TWS GRACE records, provide predictive tools to address environmental and hydrological problems, and help bridge the current gap between GRACE missions.

Climate–groundwater dynamics inferred from GRACE and the role of hydraulic memory

“…Initial investigations also established that only relatively weak first-order correlations exist between TWS and other monthly observational climate data such as the self-calibrating Palmer Drought Severity Index (PDSI-sc) (Wells et al, 2004) and mean temperature anomalies (CPC GHCN/CAMS t2m analysis) (Fan and van den Dool, 2008). By a comparison with both these measures, it appeared that PCPA carried a stronger climate variability signal due to the tropical or sub-tropical location of the selected aquifers (Allan et al, 2010;Shepherd, 2014). An analysis was then conducted to test for correlations between GWS and a series of measures of precipitation.…”

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

“…Liao et al [159] used altimetry data and auto-regressive time series modeling to forecast the water level of Qinghai Lake in China, whereas Chipman [160] used multiple optical and SAR sensors to quantify water level trends in Egyptian lakes. Large-scale regional studies have been conducted by Sutanudjaja et al [161], who forecast groundwater height from ERS scatterometer time series data over Western Europe, and Ahmed et al [162] simulating terrestrial water storage with a Nonlinear Auto-Regressive with Exogenous Input (NARX) neural network for major African watersheds with a half-year lead time.…”

“…Liu et al [140] made short-term forecasts of spring phenology in North America using VIIRS and MODIS data, while Petersen [99] developed a framework for MODIS-based crop yield estimation and forecasting in every African country. Further regional studies were conducted in Europe [161,174], Africa [162], Central Asia [130], East Asia [145], and South Asia [147]. We identified no study that conducts EO-based forecasting on a global scale.…”

The Efficiency of the Sustainable Development Policy for Energy Consumption under Environmental Law in Thailand: Adapting the SEM-VARIMAX Model

“…Their study concluded that Poland and the EU states require new approaches in terms of energy management and vehicle operation management. In Africa, Ahmed et al [89] applied the ANNs model to forecast GRACE data of African watersheds and found that the model provided the most accurate forecast. Ramsauer et al [90] adapted a Factor-Augmented Vector Autoregression Model (FAVAR) with an extension of a Kalman Filter for Factors to measure the impact of monetary policy in a case study.…”