Mapping Forest in Europe by Combining Earth Observation Data and Forest Statistics

Päivinen, Risto; Lehikoinen, Mikko; Schuck, Andreas; Häme, Tuomas; Väätäinen, Seppo; Andersson, Kristin; Kennedy, Pamela J.; Folving, Sten

doi:10.1007/978-94-017-0649-0_21

Cited by 24 publications

(30 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Present-day land cover, elevation, longitude and latitude Data on the present-day land cover in the study region were obtained from the forest map of Europe compiled by the European Forest Institute (EFI). Raster maps based on a combination of satellite data (NOAA-AVHRR) and national forest-inventory statistics from 1990-2005 (Päivinen et al, 2001;Schuck et al, 2002) were downloaded from the EFI webpage (http://www.efi.int/portal/virtual_ library/information_services/mapping_services/ forest_map_of_europe). The forest maps (with proportions of coniferous-and broadleaved-forest cover) were upscaled, by averaging, from 1 km × 1 km to 1…”

Section: Anthropogenic Deforestationmentioning

confidence: 99%

Creating spatially continuous maps of past land cover from point estimates: A new statistical approach applied to pollen data

Pirzamanbein

Lindström

Poska

et al. 2014

Ecological Complexity

View full text Add to dashboard Cite

Reliable estimates of past land cover are critical for assessing potential effects of anthropogenic land-cover changes on past earth surface-climate feedbacks and landscape complexity. Fossil pollen records from lakes and bogs have provided important information on past natural and human-induced vegetation cover. However, those records provide only point estimates of past land cover, and not the spatially continuous maps at regional and sub-continental scales needed for climate modelling.We propose a set of statistical models that create spatially continuous maps of past land cover by combining two data sets: 1) pollen-based point estimates of past land cover (from the REVEALS model), and 2) spatially continuous estimates of past land cover, obtained by combining simulated potential vegetation (from LPJ-GUESS) with an anthropogenic land-cover change scenario (KK10). The proposed models rely on statistical methodology for compositional data and use Gaussian Markov Random Fields to model spatial dependencies in the data.Land-cover reconstructions are presented for three time windows in Europe: 0.05 ka, 0.2 ka, and 6 ka years before present (BP). The models are evaluated through cross-validation, deviance information criteria and by comparing the reconstruction of the 0.05 ka time window to the present-day land-cover data compiled by the European Forest Institute (EFI). For 0.05 ka, the proposed models provide reconstructions that are closer to the EFI data than either the REVEALS-or LPJ-GUESS/KK10-based estimates; thus the statistical combination of the two estimates improves the reconstruction. The reconstruction by the proposed models for 0.2 ka is also good. For 6 ka, however, the large differences between the REVEALS-and LPJ-GUESS/KK10-based estimates reduce the reliability of the proposed models. Possible reasons for the increased differences between REVEALS and LPJ-GUESS/KK10 for older time periods and further improvement of the proposed models are discussed.

show abstract

Section: Anthropogenic Deforestationmentioning

confidence: 99%

Creating spatially continuous maps of past land cover from point estimates: A new statistical approach applied to pollen data

Pirzamanbein

Lindström

Poska

et al. 2014

Ecological Complexity

View full text Add to dashboard Cite

show abstract

“…EFMAP 2002 represents the proportion of coniferous and broadleaf forests per land area in a 1 × 1 km pixel resolution. The map was derived from NOAA-AVHRR satellite imagery, and calibrated to conform to statistical information on forest area, such as national inventory reports and data from international data-collection processes of the United Nations Economic Commission for Europe/Food and Agriculture Organisation (UNECE/FAO 2000;PÄivinen et al 2001). The difference between the map and inventory statistics varies around ± 5% (higher in the coastline and small island area) (Schuck et al 2003 (Kempeneers et al 2011), which was aggregated from the original 25 m spatial resolution to 1 km, by summing up the forest area for each 1 km pixel (the proportion of forests is given).…”

Section: Corine Land Cover (Corine -Eea 1994)mentioning

confidence: 99%

Free European data on forest distribution: overview and evaluation

Trombik¹,

Hlásny²

2013

J. For. Sci.

View full text Add to dashboard Cite

A growing need for the evaluation of prospects and sustainability of forest resources calls for the availability of harmonized data on forest distribution. We described and evaluated nine datasets providing such information: Corine LandCover, four European forest maps and four tree species distribution maps. Apart from providing a condensed overview of these datasets, we focused on the match between selected forest maps and forest management plans (FMPs) of Slovakia, which can be thought of as highly accurate information on forest distribution. The degree of match between forest and species area, within 306 forest administrative districts of Slovakia, was used as an indicator of accuracy. In addition, the match between the total forest and species area in Slovakia, given by FMPs and by evaluated datasets, was addressed. We found a high degree of match for the datasets on forest distribution (R-square 0.77–0.93, depending on the dataset), as well as strong agreement in total forest area (± 5%). Both indicators are worse in the case of forest type evaluation (coniferous and broadleaved). Poor results were obtained for tree species maps, which under- or overestimated species areas by tens of per cent, although differences were highly variable among species. The obtained results are valid mainly for temperate forests.

show abstract

“…The RMSD was modified from Häme et al (2001), equation (4). (4) where a i is the area of the ith county, is the total Wisconsin area, and ir and iv denote the estimated proportion of forest land area in the ith county obtained from FIA, CDL, or NLCD (r) and FIA, CDL, or NLCD (v), respectively, for each of three pair-wise comparisons of estimates.…”

Section: Pixel-based Area Estimationmentioning

confidence: 99%

“…(which was amended in 1992) Forest Resources Assessment (Powell et al 1993, Zhu and. Päivinen et al (2001) describe a more recent effort to combine AVHRR data with European forest statistics. The utility of these forestcover-type maps depends on their classification accuracy and their correspondence with inventory-based forest assessments.…”

Section: Introductionmentioning

confidence: 99%