摘要
Forest fire is a serious disaster all over the world. The Fire Weather Index (FWI) System can be used in ap- plied forestry as a tool to investigate and manage all types of fire. Relative humidity (RH) is a very important parameter to calculate FWI. However, RH interpolated from meteorological data may not be able to provide precise and confident values for areas between far separated stations. The principal objective of this study is to provide high-resolution RH for FWI using MODIS data. The precipitable water vapor (PW) can be retrieved from MODIS using split window tech- niques. Four-year-time-series (2000-2003) of 8-day mean PW and specific humidity (Q) of Peninsular Malaysia were analyzed and the statistic expression between PW and Q was developed. The root-mean-square-error (RMSE) of Q es- timated by PW is generally less than 0.0004 and the correlation coefficient is 0.90. Based on the experiential formula between PW and Q, surface RH can be computed with combination of auxiliary data such as DEM and air temperature (Ta). The mean absolute errors of the estimated RH in Peninsular Malaysia are less than 5% compared to the measured RH and the correlation coefficient is 0.8219. It is proven to be a simple and feasible model to compute high-resolution RH using remote sensing data.
Forest fire is a serious disaster all over the world. The Fire Weather Index （FWI） System can be used in applied forestry as a tool to investigate and manage all types of fire. Relative humidity （RH） is a very important parameter to calculate FWI. However, RH interpolated from meteorological data may not be able to provide precise and confident values for areas between far separated stations. The principal objective of this study is to provide high-resolution RH for FWI using MODIS data. The precipitable water vapor （PW） can be retrieved from MODIS using split window techniques. Four-year-time-series （2000-2003） of 8-day mean PW and specific humidity （Q） of Peninsular Malaysia were analyzed and the statistic expression between PW and Q was developed. The root-mean-square-error （RMSE） of Q estimated by PW is generally less than 0.0004 and the correlation coefficient is 0.90. Based on the experiential formula between PW and Q, surface RH can be computed with combination of auxiliary data such as DEM and air temperature （Ta）. The mean absolute errors of the estimated RH in Peninsular Malaysia are less than 5% compared to the measured RH and the correlation coefficient is 0.8219. It is proven to be a simple and feasible model to compute high-resolution RH using remote sensing data.
基金
Under the auspices of the Airborne Remote Sensing (MARS) Program of Malaysia (No. KSTAS/MACRES/T/2/2004)
