Disappearing Diaras: Three Decades of Sandbar Change and Agricultural Loss in the Ganga at Patna, Bihar
Seema Mehra Parihar
*
Department of Geography, Himalayan Centre WCG, Kirori Mal College, DHSP, Institute of Eminence, University of Delhi, India.
Mohammed Baber Ali
Department of Geography, University of Delhi, India.
Asmit Soni
Kirori Mal College, University of Delhi, India.
Yuvraj Singh Sikarwar
Kirori Mal College, University of Delhi, India.
Vyom Sharma
Kirori Mal College, University of Delhi, India.
*Author to whom correspondence should be addressed.
Abstract
The Digha Ghat–Barh reach of the middle Ganga is a highly dynamic braided river system where annual monsoon-driven sandbar formation and erosion directly shape floodplain morphology, agriculture, and settlement patterns amid long-term hydrological regulation and intense human use. This study maps how mid-channel and attached sandbars i.e. chars and diaras in local usage, have changed across a 95 km reach of the Ganga between Digha Ghat and Barh, Patna district, Bihar, from 1995 to 2025, and traces what those changes have cost the farming households who cultivate them. Seven Landsat epochs at 30 m resolution were classified using a composite NDWI–BSI–NDVI method, extracting bare sandbar, vegetated bar, and active channel surfaces for pre- and post-monsoon seasons. The reach was divided into five sub-reaches of approximately 19 km each; Braiding Index (Brice, 1964) and Sinuosity Index (Friend & Sinha, 1993) were computed for each sub-reach at each epoch, and their inverse relationship tested via Pearson's correlation interpreted on the Evans (1996) scale. A structured household survey of 130 diara-dependent households recorded geomorphic crop losses over a five-year recall window. Between 1995 and 2025, bare-bar areas in the pre-monsoon months fell or shrank to a size of 6,850 ha, representing a total decrease of 18.7%, attributed to upstream water regulation of flow rates, while the amount of area contraction from the pre-monsoon to post-monsoon time frames (i.e., areas at low water levels relative to high water levels) remained relatively stable, ranging between 35-39% for all epochs studied. A correlation between the braided index (BI) and the stacked index (SI) was negative across all epochs (pooled r = -0.68, p < .01), with Sub-Reach 3 exhibiting the most braiding and morphodynamically active of the four study segments and the most consistent source of Punpun sediment inflow to this area; while Sub-Reach 4 had a regression-bar fraction of approximately 30% due to the unregulated extraction of sand. According to results from the household surveys conducted, 72% of the households within the diara areas affected by this study had experienced at least one instance of geomorphological crop loss within the five year period preceding the time that the surveys were conducted and that the percentage of surveyed households experiencing crop loss increased from SR-1 (i.e., 52%) to SR-3 (i.e., 84%), a spatial pattern that follows the intensity of braiding observed within each of the sub-reaches studied and represents a measureable risk profile for the establishment of crop insurance; as well as the creation of infrastructure to facilitate evacuation; and for the regulation of sand mining within the mid- and lower Ganges River Basins.
Keywords: Ganga River, sandbars, chars, diaras, braiding index, sinuosity index, NDWI, NDVI, Patna, Bihar, remote sensing, sand mining, agricultural loss, fluvial geomorphology