Urbanization is considered as the key driver for land use and land cover (LULC) changes across the globe and Delhi is no exception to this phenomenon. The population of Delhi has almost doubled from 8.4 million in 1991 to 16.3 million in 2011. Correspondingly, the built-up area has also increased from 336.82 to 598.22 km 2 during 1999–2018. This urban expansion has led to emergence of serious ecological risk and fragmentation of the landscape. In this context, it is imperative to analyse the level of risks induced by such urban expansion and its underlying associations with other factors. This article quantifies the LULC changes in Delhi during 1999–2018 using Landsat 5 (TM) and Landsat 8 (OLI) data. A spatio-temporal sprawl induced risk assessment index has been developed by combining landscape fragmentation score and land use land cover vulnerability score. The landscape fragmentation score was based on four landscape metrics, whereas the vulnerability score was computed from LULC data. The article also assesses the association between risk areas and economic activities, environmental and infrastructural amenities that are considered key drivers of urban expansion in Delhi. To estimate spatio-temporal variability between risk areas and key drivers, ordinary least square regression and geographical weighted regression (GWR) were used. The GWR results reveal that sprawl-induced ecological risk in Delhi is strongly associated with economic activity, infrastructural accessibility and environmental amenities. This spatial empirical assessment also shows that urban growth incentives or services such as roads, metro rail, schools and hospitals can also create pressure on the landscape if local authorities arbitrarily provide these services across space without considering the associated risks.

Abstract. Understanding the land use and land cover (LULC) change processes is of great importance for formulating the land use policy. In the present study, Remote Sensing, Geographic Information System (GIS) and statistical techniques are used to understand the LULC dynamics and possible driving forces in a thickly populated Yamuna river basin of India. The National Capital Territory of Delhi, one of the world's largest urban agglomerations, occurs in this river basin. The LULC maps on 1:250,000 scales prepared at decadal interval (1985, 1995 and 2005) using satellite images are used for studying the LULC dynamics. The major LULC changes observed during the study period are increase in cropland (65.7% to 67.9%) and built-up (0.9% to 1.9%) and decrease in forest (17.1% to 16.4%) and water bodies (2.3% to 1.9%). The dominant processes of the LULC changes include transitions from forest, grassland, shrub land, water bodies and wasteland to cropland; transitions from cropland, shrub land, wasteland, grassland and forest to built-up; conversion of forest to cropland, shrub land, plantation and built-up; and transitions from water bodies to cropland, shrub land and wasteland. Most of the transitions are reversible in nature. The statistical analysis (logistic regression) between decadal LULC maps and drivers indicates that LULC changes are driven by a multitude of factors in different combinations and with different importance. Socioeconomic, demographic and proximity drivers mainly influence the LULC change processes in the study area in decreasing order of importance. The insights gained are important for simulating future LULC scenarios as input for land management policy making.