Land use change after the collapse of the Soviet Union

and its effects on protected areas and large mammal populations Eugenia V. Bragina1,2, A. Sieber3, T. Kuemmerle3, P. Potapov4, S. Turubanova4, A. Tyukavina4, M. Hansen4, M. Baumann1,

A.M. Pidgeon1, K.J. Wendland5, A.V. Prishchepov6, A.R. Ives7, N. Keuler8, L.M. Baskin9, V.C. Radeloff1

Protected area effectiveness

How effective were Russian protected areas during economic downturns given lack

of funding and low enforcement levels?

Goal: Evaluation of the effectiveness of Russian protected areas before and after

collapse of the socialism through changes in forest cover

Forest disturbance in European Russia

Logging in boreal and temperate forests is a major source of carbon emissions, but

region-wide forest disturbance information is lacking.

Goal: Map region-wide Forest cover and disturbance with an automatic Landsat

data processing, compositing and classification algorithm

• Location: Oksky and Mordovsky strictly protected state

nature reserves

• Data: Landsat footprints 174-175-176/22; 1985-2010

• Support Vector Machine classifications

• Trajectory analysis of forest disturbance index (DI)

(Healey et al. 2005)

disturbance index < 2 forest (undisturbed across

time)

disturbance index ≥ specific threshold non-forest

(forest disturbance)

Annual forest disturbance rates (

% of 1984 forest)

Stable forest Forest disturbance in

year 12

Approach Results

• About 40% of the 1988 farmland abandoned by 2010

• Forest disturbance inside Oksky 1.81%, outside

6.15%; inside Mordovsky 0.5%, outside 5.21%

• Lower forest disturbance rate in the 1990s (Sieber et

al. 2013)

Forest disturbance detection

Approach Results

• Location: Northern

European Russia

• Data: USGS archive of Landsat TM/+ETM

• Forest disturbance detection in 1990-2000

(Yaroshenko et al., 2008):

Per-image change detection using SWIR

band difference threshold and

unsupervised clustering

• Forest disturbance detection in 2000-2005

(Potapov et al., 2011):

Forest cover and change mapping using

supervised classification tree algorithm

and cloud-free image composites for

2000 and 2005

• Forest disturbance detection in 2000-2012:

Annual forest cover change mapping

using multi-temporal spectral metrics and

annual cloud-free image composites

• Logging contributes 89% of total forest cover loss

• The most populated regions had the highest rates of forest loss

• Between 1990-2000 and 2000-2005:

Annual gross forest cover loss decreased from 528

to 402 ha*1000/yr

Annual logging area decreased by 33%

Logging increased within the Central and Western parts

Annual burned forest area increased more than 50%

• Work in progress:

forest cover change

for 2000-2012 for all

of Eastern Europe

Annual gross forest cover loss

attributed to logging

Net forest cover change 2000–2005, as % of

forest cover for 2000

1990-2000

2000-2005

>1%

-1% – 1%

-2% – -1%

< -2%

Potapov et al. 2012

Tree canopy cover for 2000 Gross forest cover gain Gross forest cover loss 2000 -2012

Approach Results

• Location: Caucasus strictly protected state nature reserve

• Data: Landsat footprints 173/29-30; 1985-2010

• Support Vector Machine classification

• Four classes: forest, agriculture, grassland, and meadows

• Post-classification comparison. Forest either converted to

(1) agriculture, (2) grassland, or was (3) disturbed

Caucasus NR

• Forest disturbances inside nature reserve was 7.03%

• Forest disturbances outside nature reserve was 12.3%

Agriculture

Forest

Grassland

Other

Forest disturbance

Affiliations

1 1 Department of Forest and Wildlife Ecology, University of Wisconsin-

Madison, WI, USA

2 Lomonosov Moscow State University, Russia

3 Geography Department, Humboldt-Universität zu Berlin, Germany

4 Department of Geographical Sciences, University of Maryland, MD, USA

5 Department of Conservation Social Sciences, University of Idaho, ID, USA

6 Leibniz Institute for Agricultural Development in Central and Eastern

Europe, Germany

7 Department of Zoology, University of Wisconsin-Madison, WI, USA

8 Department of Statistics, University of Wisconsin-Madison, WI, USA

9 Severtsov Institute of Ecology and Evolution, Russia We gratefully acknowledge support by NASA’s Land Cover and Land Use Science Program

Large mammals population trends after the collapse of the Soviet Union

How do socioeconomic shocks affect wildlife? Quantitative evidence is sparse.

Goal: Examination of population trends of large mammal species in Russia before

and after the collapse of socialism

Approach Results

• Winter Track Count data: annual survey of animal tracks

since 1964 on up to 50,000 transects per year

• We analyzed wildlife population densities for all of Russia

and for each administrative regions

• Models N(t+1) = a + b*N(t) for each region

• Positive residuals indicate population increases

• Negative residuals indicate population declines

Wild boar populations in Pskov

region (left), and fitted model with

residuals for 1980s (black), ‘90s

(red) and ‘00s (green)

1980 1990 2000 2010

1.5

2.0

2.5

N(t)

1.5 2.0 2.5

1.5

2.0

2.5

N(t)

N(t+

1)

• Most large mammal populations declined very

rapidly after the collapse of socialism in 1991

• Only wolf increased after 1991

• After 2000 most species rebounded

Wildlife populations

trends in Russia relative

to 1991 populations

Contact information

Eugenia Bragina

Phone: 608-265-9219

E-mail: bragina@wisc.edu

http://silvis.forest.wisc.edu

0 300 km

0 25 50 km