Farmland abandonment and its determinants in the different ecological villages of the Koshi river basin, central Himalayas: Synergy of high-resolution remote sensing and social surveys.

Globally, there is a noticeable increasing trend in farmland abandonment, which directly affects farmers' livelihood and food security. The abandonment status, its determinants and impact vary by spatial and socioeconomic context. In order to study this important issue, we selected three different villages representing three ecological regions (Mountain, Hill, and Tarai) in the Koshi River Basin (KRB), and applied two methods: the Unmanned Aerial Vehicle (UAV) and a social survey. The UAV captured 3711 images and we carried out 162 households' survey with structured questionnaires. Pix4Dmapper and ArcGIS tools were used for combining and processing the images. On-screen digitalization and binary logistic regression (BLR) were applied to examine the status and determinants of farmland abandonment. The results show a higher proportion of farmland abandonment in the villages located in the Hill and Mountain regions compared to those in the Tarai region. Almost 10.3% area of total land and 22.3% area of total farmland was abandoned in the Hill village. The Tarai village had the least farm abandonment (3.7%). Farmers perceived that climate change (less precipitation, increasing temperatures, and drought), shifting occupations, crops damaged by wildlife, migration, lack of irrigation, and a labor shortage are the leading determinants of farmland abandonment. These factors varied slightly across the different ecological regions. The BLR model was a good fit with Nagelkerke's R = 0.776, with a correct model prediction (87.7%) and p = 0.032. The results from the regression model suggest that an increase in temperature (p = 0.000), decrease in rainfall (p = 0.001), lack of machinery used for farm-work (p = 0.000), lack of irrigation (p = 0.000), and reduction of labor-force (p = 0.000) are the main contributing determinants of farmland abandonment. This synergy of high-resolution remote sensing and farmers' perception-based findings facilitates the improvement of land-use governmental policies to improve farmers' quality of life and build sustainable farmland management.