Hoekman D.H.,Wageningen University |
Vissers M.A.M.,SarVision |
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing | Year: 2010
This paper describes the operational radar mapping processing chain developed and steps taken to produce a provisional wide-area PALSAR forest and land cover map covering Borneo for the year 2007, compliant with emerging international standards (CEOS guidelines, FAO LCCS). A Bayesian approach based on (unsupervised) mixture modeling followed by Markov Random Field (MRF) classification has been selected for its suitability and flexibility to deal with a situation where ground truth is sparse and sometimes ambiguous. The methodology is based on the classification of Fine Beam Single (FBS) and Fine Beam Dual (FBD) polarization (path) image pairs. To cover Borneo the equivalent of 554 standard images is required. Qualitative and quantitative validation results and findings are reported. The final overall accuracy assessment result shows the demonstration map product is in 85.5% full agreement with the independent reference dataset and in 7.8% 'partial agreement'. The accuracy achieved is widely considered adequate, a very promising result for a sub-continental high resolution map based on just single-year radar data. Approaches for further improvement of the accuracy of less accurately classified thematic classes such as grassland, cropland and shrubland are suggested. This work has been undertaken in part within the framework of the ALOS Kyoto & Carbon Initiative. © 2010 IEEE.
Gaveau D.L.A.,Center for International Forestry Research |
Sloan S.,James Cook University |
Molidena E.,Center for International Forestry Research |
Yaen H.,Center for International Forestry Research |
And 9 more authors.
PLoS ONE | Year: 2014
The native forests of Borneo have been impacted by selective logging, fire, and conversion to plantations at unprecedented scales since industrial-scale extractive industries began in the early 1970s. There is no island-wide documentation of forest clearance or logging since the 1970s. This creates an information gap for conservation planning, especially with regard to selectively logged forests that maintain high conservation potential. Analysing LANDSAT images, we estimate that 75.7% (558,060 km2) of Borneo's area (737,188 km2) was forested around 1973. Based upon a forest cover map for 2010 derived using ALOS-PALSAR and visually reviewing LANDSAT images, we estimate that the 1973 forest area had declined by 168,493 km2 (30.2%) in 2010. The highest losses were recorded in Sabah and Kalimantan with 39.5% and 30.7% of their total forest area in 1973 becoming non-forest in 2010, and the lowest in Brunei and Sarawak (8.4%, and 23.1%). We estimate that the combined area planted in industrial oil palm and timber plantations in 2010 was 75,480 km2, representing 10% of Borneo. We mapped 271,819 km of primary logging roads that were created between 1973 and 2010. The greatest density of logging roads was found in Sarawak, at 0.89 km km-2, and the lowest density in Brunei, at 0.18 km km-2. Analyzing MODIS-based tree cover maps, we estimate that logging operated within 700 m of primary logging roads. Using this distance, we estimate that 266,257 km2 of 1973 forest cover has been logged. With 389,566 km2 (52.8%) of the island remaining forested, of which 209,649 km2 remains intact. There is still hope for biodiversity conservation in Borneo. Protecting logged forests from fire and conversion to plantations is an urgent priority for reducing rates of deforestation in Borneo. © 2014 Gaveau et al.
Avitabile V.,Wageningen University |
Herold M.,Wageningen University |
Heuvelink G.B.M.,Wageningen University |
Lewis S.L.,University of Leeds |
And 33 more authors.
Global Change Biology | Year: 2016
We combined two existing datasets of vegetation aboveground biomass (AGB) (Proceedings of the National Academy of Sciences of the United States of America, 108, 2011, 9899; Nature Climate Change, 2, 2012, 182) into a pan-tropical AGB map at 1-km resolution using an independent reference dataset of field observations and locally calibrated high-resolution biomass maps, harmonized and upscaled to 14 477 1-km AGB estimates. Our data fusion approach uses bias removal and weighted linear averaging that incorporates and spatializes the biomass patterns indicated by the reference data. The method was applied independently in areas (strata) with homogeneous error patterns of the input (Saatchi and Baccini) maps, which were estimated from the reference data and additional covariates. Based on the fused map, we estimated AGB stock for the tropics (23.4 N-23.4 S) of 375 Pg dry mass, 9-18% lower than the Saatchi and Baccini estimates. The fused map also showed differing spatial patterns of AGB over large areas, with higher AGB density in the dense forest areas in the Congo basin, Eastern Amazon and South-East Asia, and lower values in Central America and in most dry vegetation areas of Africa than either of the input maps. The validation exercise, based on 2118 estimates from the reference dataset not used in the fusion process, showed that the fused map had a RMSE 15-21% lower than that of the input maps and, most importantly, nearly unbiased estimates (mean bias 5 Mg dry mass ha-1 vs. 21 and 28 Mg ha-1 for the input maps). The fusion method can be applied at any scale including the policy-relevant national level, where it can provide improved biomass estimates by integrating existing regional biomass maps as input maps and additional, country-specific reference datasets. © 2016 John Wiley & Sons Ltd.