Search

Aim of study The following paper presents an inter-comparison of three global products MCD45A1 (MODISMODerate resolution Imaging Spectrometer-Burned Area Product), L3JRC (Terrestrial Ecosystem Monitoring Global Burnt Area Product), and GLOBCARBON Burnt Area Estimate (BAE) Product; and three local products, two of them based on MODIS data and the other one based on MERIS (MEdium Resolution Imaging Spectrometer) data. Area of study The study was applied to the Galician forest fires occurred in 2006. Material and methods Materials used involved the three already mentioned global products together with two MODIS and one MERIS reflectance images, and MODIS thermal anomalies. The algorithm we used, which is based on the determination of thresholds values on infrared bands, allowed the identification of burned pixels. The determination of such threshold values was based on the maximum spatial correlation between MODIS thermal anomalies, and infrared reflectance values. This methodology was applied to MODIS and MERIS reflectance bands, and to the NBR (Normalized Burn Ratio). Burned area validation was evaluated using burned area polygons as derived from an AWiFS (Advanced Wide Field Sensor) image of 60 m pixel size. Main results Best results were reached when using the MERIS infrared bands, followed by the MODIS infrared bands. Worst results were reached when using the MCD45A1 product, which clearly overestimated; and when using the L3JRC product, which clearly underestimated. Research highlights Since the efficiency of the performance of the available burned area products is highly variable, much work is needed in terms of comparison among the available sensors, the burned area mapping algorithms and the resulting products.

This paper analyses the spatial structure of the Pinus pinaster Aiton stands within Madrid province (Central Spain) by means of high spatial resolution images, as well as the relationship between texture indexes obtained from variograms, and dasometric features of forest stands. In order to carry out this research, some orthophotographs and an IKONOS satellite image were used. All these images corresponded to locations of the Spanish National Forest Inventory (NFI) sample plots for the species of interest. Image analysis was based on geostatistic tools and particularly on the variogram. Experimental variograms were calculated using the different data sets. The following step consisted on fitting mathematical functions to them in order to produce ideal variograms. These models were used to extract a set of variogram variables that might be useful for characterizing forest stands at different spatial resolutions. As a result, some significant relationships between variogram parameters and dasometric variables were found, but of limited predictive power. The latter reveals the diff iculties in applying such methodology when working on complex topographies and spatially heterogeneous forests as the Mediterranean.

This study presents a rapid, accurate and economic methodology for forest fire burned areas estimation using remote sensing techniques. The algorithm used is based on thresholds identifications from infrared spectral band, using MODIS images and hotspots time series delivered from NASA. The threshold is defined by the maximum correlation between the infrared band and hotspot influence area. The results show the usefulness of the presented methodology in forest fire burned area estimation and its relevance for prioritizing forest management activities and for planning future reforestations work in the affected area.