Please use this identifier to cite or link to this item:
Title: Modeling of Land Surface Temperature (LST) and Normalized Difference Vegetation Index (NDVI) in Nepal: 2000-2015
Authors: Phattrawan, Tongkumchum
Ira, Sharma
Faculty of Sciecnce and Technology (Mathematics and Computer Science)
คณะวิทยาศาสตร์และเทคโนโลยี ภาควิชาคณิตศาสตร์และวิทยาการคอมพิวเตอร์
Keywords: Analysis and Modelling
Issue Date: 2018
Publisher: Prince of Songkla University, Pattani Campus
Abstract: The study explores the seasonal pattern and trend of LST and NDVI in Nepal from 2000 to 2015. The time series data of temperature (LST) and vegetation (NDVI) were derived from MODIS website. Natural cubic spline function, polynomial and logistic regression models were used to analyze LST while, natural cubic spline function, linear regression and GEE were used in case of NDVI. The data were seasonally adjusted in both LST and NDVI. For handling autocorrelation, autoregression of first order (AR (1)) was applied and filtered the data in LST and GEE model was used during NDVI analysis for this purpose. Here, LST and NDVI represented two parts of the study that are summarized as given below. The first part involves the temperature change pattern in Nepal. An area of 11,902 km2 within latitudes 26.92°N -28.26°N and longitudes 85.20°E-85.58°E was selected in form of 27 regions. Every region was further divided into nine sub regions. Therefore there were 243 sub regions which were analyzed one by one. Firstly, the seasonal pattern of temperature for 15 years revealed that seasonal changes were not basically different in the sub regions. Secondly, the data were fitted to polynomial regression of second order to obtain quadratic slopes of LST. The LST slopes illustrated the local pattern of temperature change during 15 years period which were categorized into five groups: ‘Flat’ pattern (11.5% of grid area), ‘Accelerated-increasing’ pattern (25.9% of grid area), ‘Decelerated-increasing’ pattern (20.6% of grid area), ‘Decelerated-decreasing’ pattern (22.6% of grid area) and ‘Accelerated-decreasing’ pattern (19.3% of grid area). The patterns were regrouped into binary (accelerating and non accelerating) to model with the altitude (3 categories) and land cover (three categories) of the regions. The results showed that accelerating pattern had negative association with the altitude and no vegetated land cover. When the results were described in terms of ecozones, the area in the temperate Mountain zone dominantly showed no change or gradual increase of LST pattern. Low populated, low vegetated, snow cladded land on northern Himalaya or alpine zone had apparently decreasing LST pattern. The southern high populated, tropical Tarai zone had dominant increasing pattern. The second part involves the vegetation change and seasonal pattern, using NDVI data, in three different regions of Nepal, east - Dhankuta (27.15°N, 87.35°E), center - Kathmandu (27.59°N, 85.39°E) and west - Surkhet (28.62°N, 81.88°E). Each region had an area of 410 km2 with 6,561 grids. The analysis was done separately for systematically selected 196 grids at each region. At first, the annual seasonal pattern showed the seasonal start (greening) was earlier in east and moved gradually to the west. The case of end of season (browning) had similar results. Also, the length of season was longer in east than westward. Secondly, NDVI linear trend for 15 years showed that, except at the center, east and west suburban regions had dominant increasing trend. Lastly, to adjust for autocorrelation and applying overall 196 grids in a single model, generalized estimating equations (GEE) were used. The CI plots after this model explained, the vegetation was increasing in Nepal during these 15 years, except in Kathmandu. The recent declining trend in Kathmandu is alarming.
Description: Thesis (Ph.D.(Research Methodology))--Prince of Songkla University, 2018
Appears in Collections:746 Thesis

Files in This Item:
File Description SizeFormat 
TC1547.pdf8.44 MBAdobe PDFView/Open

Items in PSU Knowledge Bank are protected by copyright, with all rights reserved, unless otherwise indicated.