FAO Global Forest Land-use Change 1990-2005 Summary_Report

8/3/2019 FAO Global Forest Land-use Change 1990-2005 Summary_Report

1/8

Gba fsan-s cangf 1990 2005Initial results rom a global remote sensing survey


2/8


3/8

Gba fs an-s cang f 1990 2005

3

Best-available Landsat satellite imagery was used to track change

in orest areaMost o the data was provided by the USGS Landsat Global Land Survey, which comprised

the best Landsat satellite images acquired closest to the target years o 1990, 2000

and 2005. Landsat imagery was used as the primary source o data because its global

coverage, long time series, spectral characteristics and 30 m spatial resolution made

it suitable or the detection o tree cover and change in tree cover over time. Image

samples were classiied to land cover and land use using an automated, supervised

approach. Nearly 7 million polygons were analyzed at each time interval to enable

detection o orest area, orest gains and orest losses 5 hectares or greater in size.

Forest land use and land-use change reportedLand cover is the observed biophysical properties o the land surace, whereas land use is

deined by the human activities and inputs on a given land area. This brochure summarizeschanges in orest land use. Changes in land cover are explored elsewhere.1

Treating orest as a land use is consistent with the orest deinition used in FAOs Global

Forest Resources Assessment country reports and national reports to the United Nations

Framework Convention on Climate Change (UNFCCC). Forest land use may include periods

during which the land is devoid o tree cover, or example during cycles o orest harvesting

and regeneration. In such cases, a land use is considered to be orest land use when

management or natural processes will, within a reasonable time, restore tree cover to the point

where it constitutes a orest. Figure 2 shows a simpliied example o the main steps used in

processing Landsat data or 1990 and 2000 to determine land use and changes in land use.

National experts in over 100 countries reviewed dataThe automated drat maps o the sample areas were reviewed by over 200 national experts

in 106 countries. The review process helped improve results and also helped introduce new

remote sensing tools in countries where remote sensing methods were not well developed.

1 See background material and orthcoming papers, which will be available at www.ao.org/orestry/ra/remotesensingsurvey/en/.

FIGURE 2

Determining orest land-use change rom satellite imagery

1) Satellite data 2) Classiy and label 3) Validate, check changes

Change

Gan n fs an-s

lss n fs an-s


4/84


Survey reSultS

The area in orest land use declined between 1990 and 2005The survey shows the total orest area2 in 2005 was 3.69 billion hectares, which is

30 percent o the global land area. Table 1 shows that there was a net reduction in the

global orest area between 1990 and 2005 o 72.9 million hectares.

Just over hal the worlds orests are in tropical or subtropical climaticdomainsThe spatial location o the samples enables the results to be presented by six regions

(Figures 3 and 4) and by ive major climatic domains (Figure 5). The worlds orests are

distributed unevenly with just under hal the worlds orests in the tropical domain (44% o

total area), about one third in boreal (34%) and smaller amounts in temperate (13%) and

subtropical (9%) domains (Figure 6).

Forest loss and gains estimated at a global scaleWorldwide, the gross reduction in orest land use caused by deorestation and natural

disasters (14.2 million hectares per year between 1990 and 2000 and 15.2 million hectares

per year between 2000 and 2005) was partially oset by gains in orest area through

aorestation and natural orest expansion (10.1 million hectares per year between 1990

and 2000 and 8.8 million hectares per year between 2000 and 2005). This resulted in net

losses o 72.9 million hectares o orests over 15 years between 1990 and 2005.

The annual net conversion rom orest land use to other land uses averaged

4.9 million hectares per year over the 15-year period. The rate o net orest loss increased

rom 4.1 to 6.4 million hectares per year between 1990 to 2000 and 2000 to 2005,

respectively (Table 2).

There were important regional dierences in orest loss and gainFigure 7 shows notable regional dierences in the rate o change in orest area. The

highest rate o orest conversion to other land uses in both periods was in South America,

ollowed by Arica. Asia was the only region to show net gains in orest area in both

periods consistent with the extensive planting that has been reported by several countries

to be larger in area than the regional losses in area o natural orests.

TABLE 1

Global orest area (1 000 ha), 1990, 2000 and 2005

Mean Lower limit* Upper Limit*

1990 3 767 000 3 683 000 3 850 000

2000 3 726 000 3 643 000 3 808 000

2005 3 694 000 3 611 000 3 776 000

* Note: Lower and upper limit represent the 95% confdence interval3.

2 The estimates and discussion reer to changes in the area o orest land use, but or convenience the term land use is omitted.3 A confdence interval indicates the range o values within which the average may lie based on extrapolating the sample

to the population and does not include other potential sources o error.


5/8


5

TABLE 2

Global orest area change (1 000 ha/yr), 19902000 and 20002005

Mean Lower limit* Upper Limit*

19902000 -4 100 -5 300 -2 900

20002005 -6 400 -8 000 -4 700

* Note: Lower and upper limit represent the 95% confdence interval.The total orest area used in these tables is based on the 13 057 samples representing the land area with good sample dataand excludes Antarctica, and the polar ecological zone where there are no or ew trees.

Forest loss was highest in the tropics and there were slight net gainsin other climatic domainsNet orest loss was highest in the tropical ecological domain in both time periods

(Figure 8), going rom 6.3 million hectares per year in the 1990s to 8.0 million hectares per

year between 2000 and 2005. Three climatic domains subtropical, temperate and boreal

appeared to have had slight net increases in orest area in both periods, although these

were not always signiicant at the 95% conidence interval level.

FIGURE 3

Map o the six regions used in this survey

FIGURE 5

Map o the fve broad climatic domains

FIGURE 6

Distribution o global orest area (2005) byclimatic domain

tca44%

Ba34%

ta13%

Sbca9%

FIGURE 4

Distribution o global orest area (2005) by region

ocana3%

e29%

S Aca

22%

N anCna Aca

19%

Asa14%

Afca13%


6/86


Boreal

199000

200005

Temperate

199000

200005

Subtropical

199000

200005

Tropical

199000

200005

World

199000

200005

FIGURE 7

Annual change in orest land-use area (19902000 and 20002005) by region

Arica

Note: Note the net result (blue) is the result o the sum o the gross gains minus losses. A negative net result indicatesoverall orest conversion to other land uses.

FIGURE 8

Annual change in orest area (19902000 and 20002005) by climatic domain

Note: Note the net result (blue) is the result o the sum o the gross gains minus losses. A negative net result indicatesoverall orest conversion to other land uses.

Gss gan

Gss ss

N cang

6000

4000

2000

0

-2000

-4000

-6000

199000

200005

Asia

199000

200005

Europe

199000

200005

North andCentralAmerica

199000

200005

Oceania

199000

200005

SouthAmerica

199000

200005

Thousandh

ectares

(000

ha)

Gss gan

Gss ss

N cang

12000

10000

8000

60004000

2000

0

-2000

-4000

-6000

-8000

-10000

-12000

-14000

-16000

-18000

Thousandh

ectares

(000

ha)


7/8


7

CompAriNG the remote SeNSiNG SurveyANd the GloBAl ForeSt reSourCeS ASSeSSmeNt2010 reSultS For ForeSt AreA

The remote sensing survey estimates o global orest land use and changes are dierent

rom the FAO Global Forest Resources Assessment 2010 (FRA 2010) estimates. This is

because they come rom two very dierent sources the FRA reporting comes rom

compiling country reports and the numbers reported in this document rom remote

sensing that is more consistent globally and across the three time periods. The FAO

Global Forest Resources Assessment (FRA 2010) estimated the total global orest area

in 2005 at 4.06 billion hectares, compared with the remote sensing survey estimate o

3.69 billion hectares, a dierence o about 9 percent, which is a good result consideringthe dierences in methods.

On the other hand, the FRA 2010 reported a substantially higher net loss in orest area

rom 1990-2005 (107.4 million ha) than ound in the remote sensing survey (72.9 million

ha) or a dierence o 32 percent. FRA 2010 suggested that the rate o net orest loss

decreased in the period 20002005, while the remote sensing survey detected an increase

in rate. The largest dierence between the two assessments was in Arica where most

countries still have very old data and were most oten rom orecasts, which reduces the

accuracy o change estimates. Furthermore, neither assessment approach is suited to

accurate assessment o tree cover at lower canopy cover densities (e.g. between 10 and

30 percent canopy cover). This introduces some uncertainty in the accuracy o numbers

particularly in dry regions and or degraded orests.

Overall, or changes in orest land use area, the use o remote sensing has the advantage

o consistency in data and methodology and this survey has been speciically designed to

detect and report on orest area changes at a global, regional and ecological zone scale.

It complements the country-based FRA process, which collects a wide range o other

inormation that is also needed to monitor and report on the worlds orests. Importantly,

the remote sensing survey provides inormation on changes in orest land use in dierent

climatic domains - something which is currently impossible to obtain rom a compilation o

national data.

The methods developed through the remote sensing survey will be used to improve the

measurement and reporting o orest area and change in orest area over time as part o the

continual improvement o the FRA process and can be adapted or use at national scales.

These results can be an important input to national and international reporting processes

where orest area and change statistics are needed such as the Convention or Biological

Diversity and the emerging initiative or Reducing Emissions rom Deorestation and Forest

Degradation in Developing countries (REDD+) under the UN Framework Convention on

Climate Change.


8/8

Global Forest Resources Assessment

F nfan n FrA, as g

www.fa.g/fs/fa cnac [email protected]

m nfan n snsng s, ncng f ,

s aaab a www.fa.g/fs/fa/snsngs/n/.

t cnas f snsng s gaf acknwg

fnanca cnbns f ean Cssn, NASA,

hnz Cn, gnns f Asaa, Fnan an Fanc, an FAo.

t ws xss n can n n wa b akn fc ffca

nn f an f s gnns ganzans.

projeCt pArtNerS

FAO Global Forest Land-use Change 1990-2005 Summary_Report

Documents

Transcript of FAO Global Forest Land-use Change 1990-2005 Summary_Report