LST variability and population growth in district of Rawalpindi, Pakistan during 1993–2018: A regional climate model based bias correction approach for LST

Abstract

Variability of satellite-based land surface temperature (LST) due to variations in vegetation cover is analyzed using the Landsat Thematic Mapper/Enhanced Thematic Mapper Plus satellite data in district of Rawalpindi, Pakistan during the period of 1993–2018. Analysis was performed over two Tehsils (sub-districts) of district Rawalpindi: Taxila and Rawalpindi. Trends and variability in LST are examined using the probability density functions (PDFs). The 10th and 90th percentile (P10 and P90) of PDFs were computed to assess the lower and upper LST percentile differences for both Tehsils. It is noted that ΔP90(2018–1993) > ΔP10(2018–1993), implying more warming in higher temperatures as compared to that in lower temperatures during the study period for Tehsil Taxila only. The applicability of a regional climate model (RCM) for bias correction of LST is discussed. A mean and variance-based bias correction was performed by using a RCM Providing REgional Climates for Impacts Studies (PRECIS). Before correction, the LST overestimation approaches to −6.48 °C for all years for district Rawalpindi except for the year 1993, however bias corrected overestimation reduces to −4.40 °C. Relationship of census data with LST over Tehsil Rawalpindi is somewhat stronger using PRECIS based bias correction. The correlation between LST and Green Vegetation Fraction (GVF) is −0.55; the negative sign indicates a decrease in vegetation cover resulting in an increase of LST.