Main content area

Dead fuel moisture estimation with MSG-SEVIRI data. Retrieval of meteorological data for the calculation of the equilibrium moisture content

Nieto, Héctor, Aguado, Inmaculada, Chuvieco, Emilio, Sandholt, Inge
Agricultural and forest meteorology 2010 v.150 no.7-8 pp. 861-870
water content, fire hazard, remote sensing, meteorological data, estimation, air temperature, relative humidity, satellites, spectral analysis, vegetation, precipitation, accuracy, prediction, Spain
In this study we propose to use remote sensing data to estimate hourly meteorological data and then assess the moisture content of dead fuels. Three different models to estimate the equilibrium moisture content (EMC) were applied together with remotely sensed retrieved air temperature and relative humidity. The input data were acquired by the Spinning Enhanced Visible and Infrared Imager (SEVIRI) sensor, on board the Meteosat Second Generation (MSG) satellite, from which air temperature and relative humidity were estimated every 15min. Air temperature estimations are based on the Temperature-Vegetation Index (TVX) algorithm. This algorithm exploits the inverse linear relationship between the land surface temperature and the vegetation fractional cover. This relationship was evaluated in a spatial window where the meteorological forcing is assumed to be constant. To estimate the vapour pressure, a linear relationship between precipitable water content and vapour pressure has been derived. Precipitable water content was estimated with the thermal infrared bands of SEVIRI using a split-window algorithm and data from ground meteorological stations in Spain during the year 2005 were used to calibrate and validate the vapour pressure models. Finally air temperature and vapour pressure were combined to calculate the EMC for dead fuels and the transfer of errors of these estimates have been assessed with ground meteorological data for three different EMC models. Promising results were obtained, with mean absolute errors ranging from 1.9% to 2.7% of moisture content depending on the applied EMC model, but the remote sensed EMC tends to underestimate the EMC from ground data. Improvements in air temperature and vapour pressure estimations would lead to a better agreement between the observed and the predicted values.