Final Activity Report Summary - FORCE (Forest Changes Inventory for Climate Modelling)
The project aimed to develop with the use of Earth observation techniques the methods to collect up-to-date information regarding the geographical extension and characteristics of natural and human inducted disturbances in the boreal forest of the entire Eurasian continent as a contribution to greenhouse gases fluxes quantification to support the UN FCCC and Kyoto protocol.
The FORCE project exploited time-series of moderate resolution data from number of the satellite instruments (SPOT-Vegetation, Terra-MODIS, Envisat-MERIS) combined with high-resolution images (Landsat-TM/ETM+, SPOT-HRV/HRVIR, MSU-E/Meteor-3M) and in-situ information for continuous assessment of forest changes caused by fires, logging and industrial pollution.
The burnt area mapping method is based on inter-annual forest change detection from SPOT-Vegetation time-series data for consecutive years combined with active fires detected with MODIS. The burnt area database covering Northern Eurasia for the period 2000-2007 has been derived using the developed method. Three different methods have been evaluated to estimate forest burnt severity. One of these methods is based on spectral-temporal mixture analysis with use of pre- and post-fire SPOT-Vegetation data. The benefit of MERIS data for burnt severity assessment has been investigated based on MTCI index, which demonstrated strong correlation with burnt severity and was found to be able to classify up to five levels of trees mortality.
A combination of multi-temporal high-resolution satellite data (e.g. MSU-E/Meteor-3M and Landsat-TM/ETM+) acquired before and after fire with sampling of ground reference data has been applied as well for burnt severity assessment and demonstrated high level of consistency. The site-level burnt severity estimates have been used as reference to derive regional assessment through up-scaling using SPOT-Vegetation data derived products. The forest fire carbon emission model, which requires burnt area and severity estimated together with fuel material types data has been applied to produce carbon emission assessment.
The study demonstrated the possibility to use 250-m spatial resolution MODIS data to map logging area. The developed method is based on cloud-free weekly MODIS composites and has been tested over European part of Russia and Central Siberia. All developed MODIS data processing methods allows use in fully automatic mode over large region and can be extended at continental level of Northern Eurasia. The logging detection accuracy with MODIS data has been estimated using reference data derived from high-resolution imagery. The possibility to estimate carbon emissions from logging with use of satellite data has been evaluated. A combination of satellite data derived products with sampling of forest inventory data have been used to retrieve relative stock, species composition and age of forest harvested. The retrieved estimates have been evaluated in comparison to official statistics on logging and have demonstrated high level of correspondence.
The possibility to quantify forest degradation due to industrial pollution with remote sensing data has been investigated in Norilsk test region. The benefit to use both SPOT-Vegetation and MERIS data for the characterization of forest degradation caused by pollutions has been investigated. Analysis of SPOT-Vegetation data for the period 1998-2005 demonstrated a negative NDVI trend in the Norilsk surrounding, which corresponds to the dominating wind directions. The MERIS data derived MTCI index has demonstrated high sensitivity to the earlier stages degradation, which were not detectable with the SPOT-Vegetation NDVI data.
The FORCE project reintegration phase was focused on arrangements of conditions for enforcement EU-Russia collaborative research activity, training of research team in the host institution of Russia and broader/operational use of developed methods and data. A new research line has been established during the reintegration phase with focus on development of land cover dynamic mapping method and land cover map of Northern Eurasia using MODIS 250-m data.
The FORCE project exploited time-series of moderate resolution data from number of the satellite instruments (SPOT-Vegetation, Terra-MODIS, Envisat-MERIS) combined with high-resolution images (Landsat-TM/ETM+, SPOT-HRV/HRVIR, MSU-E/Meteor-3M) and in-situ information for continuous assessment of forest changes caused by fires, logging and industrial pollution.
The burnt area mapping method is based on inter-annual forest change detection from SPOT-Vegetation time-series data for consecutive years combined with active fires detected with MODIS. The burnt area database covering Northern Eurasia for the period 2000-2007 has been derived using the developed method. Three different methods have been evaluated to estimate forest burnt severity. One of these methods is based on spectral-temporal mixture analysis with use of pre- and post-fire SPOT-Vegetation data. The benefit of MERIS data for burnt severity assessment has been investigated based on MTCI index, which demonstrated strong correlation with burnt severity and was found to be able to classify up to five levels of trees mortality.
A combination of multi-temporal high-resolution satellite data (e.g. MSU-E/Meteor-3M and Landsat-TM/ETM+) acquired before and after fire with sampling of ground reference data has been applied as well for burnt severity assessment and demonstrated high level of consistency. The site-level burnt severity estimates have been used as reference to derive regional assessment through up-scaling using SPOT-Vegetation data derived products. The forest fire carbon emission model, which requires burnt area and severity estimated together with fuel material types data has been applied to produce carbon emission assessment.
The study demonstrated the possibility to use 250-m spatial resolution MODIS data to map logging area. The developed method is based on cloud-free weekly MODIS composites and has been tested over European part of Russia and Central Siberia. All developed MODIS data processing methods allows use in fully automatic mode over large region and can be extended at continental level of Northern Eurasia. The logging detection accuracy with MODIS data has been estimated using reference data derived from high-resolution imagery. The possibility to estimate carbon emissions from logging with use of satellite data has been evaluated. A combination of satellite data derived products with sampling of forest inventory data have been used to retrieve relative stock, species composition and age of forest harvested. The retrieved estimates have been evaluated in comparison to official statistics on logging and have demonstrated high level of correspondence.
The possibility to quantify forest degradation due to industrial pollution with remote sensing data has been investigated in Norilsk test region. The benefit to use both SPOT-Vegetation and MERIS data for the characterization of forest degradation caused by pollutions has been investigated. Analysis of SPOT-Vegetation data for the period 1998-2005 demonstrated a negative NDVI trend in the Norilsk surrounding, which corresponds to the dominating wind directions. The MERIS data derived MTCI index has demonstrated high sensitivity to the earlier stages degradation, which were not detectable with the SPOT-Vegetation NDVI data.
The FORCE project reintegration phase was focused on arrangements of conditions for enforcement EU-Russia collaborative research activity, training of research team in the host institution of Russia and broader/operational use of developed methods and data. A new research line has been established during the reintegration phase with focus on development of land cover dynamic mapping method and land cover map of Northern Eurasia using MODIS 250-m data.