Assessment of Climate Change Impact for Water Scarcity in Semajid Watershed, Pamekasan, East Java

Halik, Gusfan
Priyanto, Agil
Wiyono, Retno Utami Agung

Journal of the Civil Engineering Forum
Vol. 12 No. 2 (May 2026)

The increase in global temperature has caused climate change, resulting in changes in the distribution of rainfall patterns, seasonal shifts, changes in water availability, and water scarcity. At present, water scarcity in Semajid watershed in Pamekasan Regency is increasing with climate change. Water scarcity will be increasingly difficult to predict due to highly complex dynamics of atmospheric circulation and local climate phenomena such as El Niño-Southern Oscillation (ENSO). This research aims to develop an assessment model to evaluate the impact of climate change on water scarcity using the Semajid watershed of Pamekasan Regency as a case study. The prediction of water scarcity is based on atmospheric circulation dynamics data from the General Circulation Model (GCM-MIROC5) under different climate change scenarios namely Representative Concentration Pathways (RCP). A statistical downscaling model was developed to overcome the limited resolution of the GCM output. The rainfall prediction model was developed using a deep learning-based downscaling model i.e. Long-Short Term Memory (LSTM), while streamflow or water availability prediction was conducted using the Soil Water Assessment Tools (SWAT) model. The Standardized Precipitation Index (SPI) and the Water Scarcity Index (WSI) were used to assess water scarcity. The results showed that the LSTM-based downscaling model provided satisfactory rainfall predictions under different climate change scenarios (RCP) with a reliability average of R2 = 0.741. The SWAT model results also provided satisfactory predictions of water availability with an average reliability of R2 = 0.668. The assessment of water scarcity using SPI and WSI indices showed that water scarcity ranged from moderate to high levels and coincided with the occurrence of El Niño events. Overall, this study demonstrates that the integration an LSTM-based rainfall downscaling model and the SWAT hydrological model can be used as an effective tool to predict water scarcity in the Semajid watershed.