ภูมิสารสนเทศเพื่อคาดการณ์พื้นที่ที่มีโอกาสเกิดไฟป่าจากความแปรปรวนของสภาพภูมิอากาศ

Abstract

A wildfire is a problem nowadays in northern Thailand by cause natural fuel noncontrolling. And, a wildfire affects ecology such as animals, plants and watershed, including indirectly affecting the health of persons from air pollution. In addition, wildfire severity each a year has different by climate variability. The study aimed to 1) to study the difference of hotspot patterns from climate variability and 2) to study the relationship of physical factors involved in wildfire, leading to wildfire opportunity predictive model from climate variability in Omkoi District Study Area, Chiang Mai Province by using satellite data in this study consists: 1) hotspot from Terra and Aqua data, MODIS sensor, were used to analyze hotspot patterns conducted by kernel, quadrate, and nearest neighbor analysis to determine the density, dispersion and cluster of hotspots, 2) Landsat 8 satellite images were used to analyze physical factors which are land surface temperature, drought, forest type, and above-ground carbon sequestration, And 3) ALOS PALSAR satellite images used to analyze physical factors which are digital elevation model and slope. The analysis of physical factors aimed to analyze the relation with hotspot by Pearson analysis to select physical factors for wildfire opportunity predictive model creation by linear regression analysis. The period of times of this study are El Niño year (2015 and 2019) normal year (2013 and 2017), and La Niña year (2011 and 2021). The study result was found that hotspot density in El Niñohas highest, followed by normal and La Niña years, respectively. Moreover, physical factors which the relation with wildfire from linear regression analysis found that the hotspot in El nino has highest, followed by normal and La Niña years. And, physical factors influencing wildfires are land surface temperature and drought factors. In additional, wildfire opportunity predictive model creation found that model in El Niño year has the highest precision followed by La Niña year, and normal year with the accuracy of 70.2, 54.6, and 50.7 percent, respectively.