Using the Landsat record to detect forest-cover changes during and after the collapse of the Soviet Union in the temperate zone of European Russia

被引：83

作者：

Baumann, Matthias ^{[1
]}

Ozdogan, Mutlu ^{[1
]}

Kuemmerle, Tobias ^{[2
,3
]}

Wendland, Kelly J. ^{[4
]}

Esipova, Elena ^{[5
]}

Radeloff, Volker C. ^{[1
]}

机构：

[1] Univ Wisconsin, Dept Forest & Wildlife Ecol, Madison, WI 53706 USA

[2] Humboldt Univ, Dept Geog, D-10099 Berlin, Germany

[3] Potsdam Inst Climate Impact Res PIK, D-14412 Potsdam, Germany

[4] Univ Idaho, Coll Nat Resources, Moscow, ID 83844 USA

[5] Transparent World, Moscow 119021, Russia

来源：

REMOTE SENSING OF ENVIRONMENT | 2012年 / 124卷

关键词：

Forest-cover change; SVM; Temperate forests; Central and Eastern Europe; Post-Soviet land-use change; Logging; Winter imagery; Landsat; Stratified random sample; SUPPORT VECTOR MACHINES; TIME-SERIES STACKS; FARMLAND ABANDONMENT; IMAGE CLASSIFICATION; DISTURBANCE; ACCURACY; RESTITUTION; ESTIMATORS; MANAGEMENT; AREA;

D O I：

10.1016/j.rse.2012.05.001

中图分类号：

X [环境科学、安全科学];

学科分类号：

08 ; 0830 ;

摘要：

The political breakdown of the Soviet Union in 1991 provides a rare case of drastic changes in social and economic conditions, and as such a great opportunity to investigate the impacts of socioeconomic changes on the rates and patterns of forest harvest and regrowth. Our goal was to characterize forest-cover changes in the temperate zone of European Russia between 1985 and 2010 in 5-year increments using a stratified random sample of 12 Landsat footprints. We used Support Vector. Machines and post-classification comparison to monitor forest area, disturbance and reforestation. Where image availability was sub-optimal, we tested whether winter images help to improve classification accuracy. Our approach yielded accurate mono-temporal maps (on average > 95% overall accuracy), and change maps (on average 93.5%). The additional use of winter imagery improved classification accuracy by about 2%. Our results suggest that Russia's temperate forests underwent substantial changes during the observed period. Overall, forested areas increased by 4.5%, but the changes in forested area varied over time: a decline in forest area between 1990 and 1995 (- 1%) was followed by an increase in overall forest area in recent years (+ 1.4%, 2005-2010), possibly caused in part by forest regrowth on abandoned farmlands. Disturbances varied greatly among administrative regions, suggesting that differences in socioeconomic conditions strongly influence disturbance rates. While portions of Russia's temperate forests experienced high disturbance rates, overall forest area is expanding. Our use of a stratified random sample of Landsat footprints, and of summer and winter images, allowed us to characterize forest dynamics across a large region over a long time period, emphasizing the value of winter imagery in the free Landsat archives, especially for study areas where data availability is limited. (c) 2012 Elsevier Inc. All rights reserved.

引用

页码：174 / 184

页数：11

共 64 条

[11] From management to mediation: Local forestry management and the forestry crisis in post-socialist Russia [J].

Eikeland, S ;

Eythorsson, E ;

Ivanova, L .

ENVIRONMENTAL MANAGEMENT, 2004, 33 (03) :285-293

[12]

FAO, 1996, Forest Resources Assessment 1990: Survey of Tropical Forest Cover & Study of Change Processes (No. 130)

[13]

FAO, 2010, 163 FAO, P378

[14] Global consequences of land use [J].

Foley, JA ;

DeFries, R ;

Asner, GP ;

Barford, C ;

Bonan, G ;

Carpenter, SR ;

Chapin, FS ;

Coe, MT ;

Daily, GC ;

Gibbs, HK ;

Helkowski, JH ;

Holloway, T ;

Howard, EA ;

Kucharik, CJ ;

Monfreda, C ;

Patz, JA ;

Prentice, IC ;

Ramankutty, N ;

Snyder, PK .

SCIENCE, 2005, 309 (5734) :570-574

[15] The use of small training sets containing mixed pixels for accurate hard image classification: Training on mixed spectral responses for classification by a SVM [J].

Foody, Giles M. ;

Mathur, Ajay .

REMOTE SENSING OF ENVIRONMENT, 2006, 103 (02) :179-189

[16] Toward intelligent training of supervised image classifications: directing training data acquisition for SVM classification [J].

Foody, GM ;

Mathur, A .

REMOTE SENSING OF ENVIRONMENT, 2004, 93 (1-2) :107-117

[17] A relative evaluation of multiclass image classification by support vector machines [J].

Foody, GM ;

Mathur, A .

IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2004, 42 (06) :1335-1343

[18] Status of land cover classification accuracy assessment [J].

Foody, GM .

REMOTE SENSING OF ENVIRONMENT, 2002, 80 (01) :185-201

[19] Stratified sampling of satellite images with a systematic grid of points [J].

Gallego, FJ .

ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING, 2005, 59 (06) :369-376

[20] Tropical forests were the primary sources of new agricultural land in the 1980s and 1990s [J].

Gibbs, H. K. ;

Ruesch, A. S. ;

Achard, F. ;

Clayton, M. K. ;

Holmgren, P. ;

Ramankutty, N. ;

Foley, J. A. .

PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA, 2010, 107 (38) :16732-16737

← 1 2 3 4 5 6 7 →