3 Comparison of Forest Area Change FRA 1990 - FRA 2000

The primary reason for undertaking a comparison of the net annual forest area change figures of FRA 1990 and FRA 2000 relating to the periods 1980-1990 and 1990-2000 respectively is to investigate if there are differences between the net annual forest area change between the 1980's and the 1990's.

The current study compares AreaChange_FRA1990 with AreaChange_FRA2000 and discusses the magnitude and the causes of observed differences.

3.1 Starting points

The figures for net annual forest area change for the period 1980-1990 (AreaChange_FRA1990) were extracted from FRA 1990 (FAO forestry paper 124 - table 4⁴). The figures for net annual forest area change for the period 1990 - 2000 (AreaChange_FRA2000) were extracted from FRA 2000 (FAO Forestry paper 140 - appendix 3).

Especially for the industrialized countries reported by UNECE/FAO the FRA definitions of forest have changed between the FRA 1990 and the FRA 2000 (see section 2.2). For developing countries the estimates for AreaChange_FRA1990 have been generated by means of a deforestation model (also referred to as forest area adjustment model) while for FRA 2000 they have, where possible, been estimated by means of extrapolation of existing datasets from two or more points in time. For FRA 2000 many countries still relied on expert opinion and secondary sources for forest area change assessment.

AreaChange_FRA1990 and AreaChange_FRA2000 are based on data, definitions and methodologies from the respective assessments only, i.e. AreaChange_FRA2000 has been derived using FRA 2000 data only and AreaChange_FRA1990 has been derived using FRA 1990 data only. Any differences between AreaChange_FRA1990 and AreaChange_FRA2000 are therefore not due to using two data sets based on different definitions of forest to calculate AreaChange_FRA2000 or AreaChange_FRA1990.

Differences owing to e.g. using different definitions and or methodologies between the two assessments are however not eliminated.

The same fundamental preconditions as mentioned for the comparison of 1990 forest areas from the two assessments, in section 2.1 apply for the comparison of forest area change estimates. In order to compare net annual forest area change figures between FRA 1990 and FRA 2000 the country must be present in both assessments. The national units compared have by necessity been the common units (e.g. where a country has been divided into several states between the FRA 1990 and the FRA 2000, the unit for comparison has been the extent of the country under FRA 1990). Using the same process as described in Chapter 2 for narrowing the number of countries, the comparison is limited to the 176 countries and territories indicated in appendix 1 & 2.

Where annual forest area change has been reported as n.s. or n.a. it has been interpreted as equalling 0 ha.

3.2. Comparison of AreaChange_FRA1990 and AreaChange_FRA2000.

The figures for AreaChange_FRA1990 and AreaChange_FRA2000 are presented in appendix 2. The differences in forest area change between the two assessments have been compared by using the area figures, rather than the relative figures. This is so, as even minor changes in forest area of small countries or countries of low forest area produce a large relative difference without necessarily reflecting a large difference in hectares. Using area figures allows for focusing on the countries that account for the most significant difference in the net annual forest area change from the 1980's to the 1990's as reported by the FRA 1990 and the FRA 2000.

In figures 3-01 AreaChange_FRA2000 has been plotted against AreaChange_FRA1990 for all 176 countries. Figure 3-02 shows the same graph with a different level of zoom causing the plots of the major contributors of change (China, Indonesia and Brazil) to be omitted. If there were no differences between the net annual forest area change between the 2 periods, the plots would be aligned along a straight line that passes through origo at a 45^o angle. The punctured lines in figures 3-01 and 3-02 describe this trend. For similar graphs for the individual regions please refer to Appendix III.

For countries where AreaChange_FRA2000 > AreaChange_FRA1990 (i.e. where the net annual forest area change of a country has increased⁵ from the 1980's to the 1990's) the country will be plotted as lying above the punctured line and will be indicated as a positive development.

For countries where AreaChange_FRA2000 < AreaChange_FRA1990 (i.e. where the net annual forest area change of a country has decreased⁶ from the 1980's to the 1990's) the country will be plotted as lying below the punctured line and is indicated as a negative development.

Only for a few small countries are AreaChange_FRA2000 = AreaChange_FRA1990.

To allow focusing on countries that show "large differences" in net annual forest area change between the two assessments, a threshold value has been defined arbitrarily as a difference between AreaChange_FRA2000 and AreaChange_FRA1990 of +/- 100,000 ha.

The margins of +/- 100.000 ha that defines the countries accounting for large differences between AreaChange_FRA2000 and AreaChange_FRA1990 have been indicated with solid lines in the below graphs. Countries that fall between these two lines are considered to have little difference between the annual forest area change from FRA 1990 and FRA 2000. Countries that fall outside the solid lines show large differences between AreaChange_FRA2000 and AreaChange_FRA1990 that exceed +/- 100.000 ha.

Defining anything below +/- 100.000 ha difference between AreaChange_FRA2000 and AreaChange_FRA1990 as minor can rightfully be discussed and the validity of this threshold value will naturally be dependent on the country in question. However for the sake of this descriptive analysis of broad global trends, and for the sake of identifying the countries representing the major differences in forest area changes in global context from the 1980's to the 1990's it has been employed.

The initial focus of the comparison is at global/regional/subregional level with various levels of zoom. Countries that show large differences between AreaChange_FRA2000 and AreaChange_FRA1990 are presented separately in table 3.

Figure 3-01 AreaChange_FRA2000 plotted against AreaChange_FRA1990 for all 176 Countries.

Figure 3-02 AreaChange_FRA2000 plotted against AreaChange_FRA1990 for all countries (excluding Brazil, China and Indonesia).

FRA 1990 figures indicate a net annual change in forest area⁷of - 13.1 million ha during the period 1980-1990. FRA 2000 figures indicate a net annual change in forest area during the period 1990-2000 of - 9.4 million ha. In total there is a difference of 3.7 million hectares between the annual net forest area change during the 1980's and the 1990's according to the FRA data. Expressed in another way, there seems to be a net decrease in the annual deforestation rate of 3.7 million ha from the 1980's to the 1990's. Table 3-01 provides a regional/subregional breakdown of the annual net forest area change figures and their difference.

FRA 2000 RSS indicated a net rate of change which was slightly lower during the 1990's than in the 1980's however the observed differences were not statistically significant for any of the tropical regions. The FRA 2000 RSS findings on the forest area changes through the 1980's and the 1990's confirm a continued high rate of forest area loss in the tropics during the 1990's. The findings of FRA 2000 RSS corresponded well with the country data for forest area change for Tropical Asia and Tropical Latin America. For Tropical Africa however, a large discrepancy between the country data and the RSS was found, making it necessary to make an adjustment of the figures for forest area change at regional level. This is discussed further in section 4.2.

3.2.1 Regional / Subregional summary of annual net forest area change FRA 1990 - FRA 2000

Table 3-01. Annual forest area change by subregion and region FRA 1990 and FRA 2000.

Subregion / Region	FRA 1990 Figures				FRA 2000 Figures				Difference in annual change 000 ha
	Total forest 1980	Total forest 1990	Forest area change 1980-1990		Total forest 1990	Total forest 2000	Forest area change 1990-2000
	000 ha	000 ha	Annual change		000 ha	000 ha	Annual change
			000 ha	%			000 ha	%
North Africa	7,099	6,906	-19	-0.3	5,930	6,262	34	0.5	53
West Africa	104,868	96,847	-802	-0.8	98,586	85,079	-1,352	-1.6	-550
East Africa	109,723	99,693	-1,003	-1.0	148,995	135,423	-1,356	-1.0	-353
Central Africa	216,011	204,815	-1,120	-0.5	236,532	228,011	-852	-0.4	268
Southern Africa	147,406	136,684	-1,072	-0.8	212,260	194,852	-1,740	-0.9	-668
Insular Africa	145	144	0	-0.1	172	212	4	1.9	4
AFRICA	585,252	545,089	-4,016	-0.7	702,475	649,839	-5,262	-0.8	-1,246
West Asia *	12,862	12,767	-20	-0.2	22,531	22,842	31	0.1	51
Central Asia **	9,406	9,406	0	0.0	11,245	10,645	-60	-0.6	-60
South Asia	72,913	77,763	485	0.6	77,643	76,664	-97	-0.1	-582
East Asia	162,308	170,418	811	0.5	183,973	202,019	1,804	0.9	993
South East Asia	247,346	219,318	-2,803	-1.3	235,203	211,914	-2,328	-1.1	475
ASIA	504,835	489,672	-1,527	-0.3	530,595	524,084	-650	-0.1	877
Australia & New Zealand	39,831	47,309	1	0.0	164,915	162,485	-243	-0.1	-244
Other Oceania	42,204	40,945	-126	-0.3	36,261	35,052	-121	-0.3	5
OCEANIA	82,035	88,254	-125	-0.1	201,176	197,537	-364	-0.2	-239
Northern Europe ***	44,494	53,246	6	0.0	57,566	57,968	41	0.1	36
Central Europe	46,314	47,553	124	0.3	50,358	51,873	153	0.3	29
Southern Europe	32,066	39,398	58	0.1	49,397	51,723	233	0.5	175
Belarus, Moldova, Other Former USSR and Ukraine ****	754,445	754,958	51	0.0	894,000	901,388	740	0.1	689
EUROPE	877,319	895,155	239	0.0	1,051,321	1,062,952	1,167	0.1	928
North America excl. Mexico	459,902	456,737	-317	-0.1	466,684	470,564	388	0.1	705
Central America & Mexico	81,337	70,287	-1,105	-1.6	82,738	73,029	-972	-1.3	133
Caribbean	4,948	4,252	-70	-1.6	5,580	5,711	14	0.2	84
N & C AMERICA	546,187	531,276	-1,491	-0.3	555,002	549,304	-570	-0.1	921
Tropical South America	909,358	849,647	-5,971	-0.7	868,702	834,142	-3,456	-0.4	2,515
Non-Tropical South America	45,421	43,282	-214	-0.5	54,029	51,476	-255	-0.5	-41
SOUTH AMERICA	954,779	892,929	-6,185	-0.7	922,731	885,618	-3,711	-0.4	2,474
Global (excluding countries mentioned in Table 2.2)	3,550,407	3,442,375	-13,105	-0.4	3,963,300	3,869,334	-9,390	-0.2	3,715

* Figures for West Asia are excluding forest area for Armenia, Azerbaijan and Georgia. These did not exist as separate countries for FRA1990 and for reasons of comparison they have been included under Other Former USSR.

** Figures for Central Asia is only for Mongolia. The remaining 5 countries of the subregion (Kazakhstan, Usbekhistan, Turkmenistan, Tadjikistan and Kyrgistan) did not exist as separate countries for FRA 1990 and are included under Other Former USSR.

*** Figures for Northern Europe is excluding forest area for Estonia, Latvia and Lithuania, that are included under Other Former USSR.

**** Figures for Other Former USSR includes the area figures for Armenia, Azerbaijan and Georgia Kazakhstan, Usbekhistan, Turkmenistan, Tadjikistan, Kyrgistan, Estonia, Latvia and Lithuania for reasons of comparability between FRA 1990 and FRA 2000.

On a regional basis the country data for Africa and Oceania show net negative differences in the net forest area change rates from the 1980's to the 1990's, meaning an apparent acceleration of the deforestation rate in these regions from the 1980's to the 1990's. The remaining regions show a positive trend meaning that the net rate of deforestation appears to have slowed down from the 1980's to the 1990's (most explicit for South America with a positive difference of about 2,5 million ha.)⁸ In the case of Europe (including former USSR) the reported rate of forest area increase has increased from the 1980's to the 1990's.

The subregional trends for the differences are as follows:

Africa

Positive differences for North Africa, Central Africa and Insular Africa. Negative differences for West, East and Southern Africa.

North Africa. The positive difference of 53,000 ha/year may be due to plantation activity (indicated for FRA 2000 at 60,000 ha per year).

Central Africa. Democratic Republic of Congo shows a positive difference between AreaChange_FRA2000 and AreaChange_FRA1990 of about 200,000 ha. The forest area change rates of FRA 1990 for the subregion may have been overestimated due to lack of reliable baseline data and AreaChange_FRA1990 being generated by the deforestation model. Another contributing factor to the decrease in the negative rates of forest area change from the 1980's to the 1990's may be the unrest and civil wars in the subregion. Plantation establishment in the subregion is indicated as 10,200 ha and can only account for a fraction of the difference in net forest area change of 268,000 ha.

Insular Africa. The positive difference of 4,000 ha/year may primarily be due to plantation establishment (indicated for FRA 2000 at 5,200 ha per year).

West Africa, East Africa and Southern Africa show negative differences in the figures for net forest area change (-550,000 ha, -353,000 ha and -668,000 ha respectively). The main countries behind these negative figures are Côte d'Ivoire (source data referenced 1987), Nigeria (referenced 1994), Sudan (referenced 1990), Zambia (referenced 1974/93) and Zimbabwe (referenced 1992).

United Republic of Tanzania accounts for a positive difference in net annual forest area change between FRA 1990 and FRA2000 of 338,000 ha. This may be attributed to new inventory data becoming available in 1995 and a possible overestimation of the forest area change rate via the deforestation model of FRA 1990.

FRA 2000 Pan Tropical Remote Sensing Survey arrives at an estimate for net annual forest area loss which is app. 3 million ha lower than indicated the country data for deforestation during the 1990's for tropical Africa. It is likely that in lack of comprehensive inventory information for several countries, the overall negative forest area change estimates reported for FRA 2000 are overestimated (e.g. Sudan and Zambia).

Asia

Positive difference for West and East Asia. Negative difference for South, Southeast and Central Asia (Mongolia).

West Asia: Shows a small positive change (51,000 ha) assumed mainly to be due to plantation establishment in low forest area countries (indicated for FRA 2000 as 88,600 ha).

Central Asia: Due to changes in national boundaries only Mongolia is included in this subregion, while Kazakhstan, Kyrgistan, Tajikistan, Turkmenistan and Uzbekistan are included in the reporting unit Other Former USSR. The annual change in forest area was reported as 0 ha for FRA 1990. The reported annual forest area change -60.000 ha is due to loss through forest fires.

South Asia: The net difference of - 582,000 ha is mainly due to a large difference between AreaChange_FRA2000 and AreaChange_FRA1990 for India (-632,000 ha). This is in spite of an annual rate of plantation establishment for India reported to FRA1990 for the 1980's as approximately 1 million ha and as approximately 1,5 million ha for the 1990's reported to FRA 2000. Difference in methodology for generating the forest area change figures between FRA1990 and FRA2000 may partly explain the difference in the reported forest area change for India. For the subregion, minor negative differences exist between AreaChange_FRA2000 and AreaChange_FRA1990 for Nepal and Sri Lanka. Bhutan, Pakistan and Bangladesh report a positive development for annual net forest area changes from the 1980's to the 1990's. In the case of Pakistan and Bangladesh increased rates of plantations establishment during the 1990's may contribute to the positive difference in AreaChange_FRA2000.

East Asia: The net positive difference between AreaChange_FRA2000 and AreaChange_FRA1990 of 993,000 ha for the subregion owes primarily to the forest area change figures from China. For China a net increase in annual forest area change of 1,066,200 ha was reported from the 1980's to the 1990's. This major increase owes to a reported AreaChange_FRA2000 of 1,806,000 ha, mainly due to a reported annual plantation establishment rate of 1,153,800 ha.

South East Asia.
AreaChange_FRA1990 and AreaChange_FRA2000 and difference between them are marginal for Brunei Darussalaam and Singapore.

Cambodia, Lao P.D.R., Malaysia, Philippines, Thailand and Vietnam show positive differences between AreaChange_FRA1990 and AreaChange_FRA2000. Contributing factors to this trend may be a combination of reduction of natural forest resources from the 1980's to the 1990's, logging bans in natural/old growth forests, improved monitoring, increased plantation establishment, inclusion of rubber plantation establishment in AreaChange_FRA2000 along with change of methodology for estimating the annual net forest area change from FRA 1990 to FRA 2000.

The figures for Indonesia and Myanmar indicate an increase in the rate of deforestation from the 1980's to the 1990's and may be attributed to intensified shifting agriculture along with gradual conversion more permanent agriculture.

Oceania

Negative difference for the subregion Australia and New Zealand. Small positive difference for Other Oceania.

Australia & New Zealand. Negative difference owes to a large negative figure for Australia. This is due to AreaChange_FRA1990 being reported as equalling +600 ha for Australia, while for FRA 2000 the AreaChange_FRA2000 for Australia was estimated and reported at -282,000 ha as a result of improved data, and changes of FRA definitions.

With the removal of agricultural subsidies in New Zealand marginal agricultural land and pastures have reverted to shrubland/forest which combined with plantations establishment has led to a positive difference between AreaChange_{FRA2000 and} AreaChange_FRA1990 (for New Zealand this means an increased forest area during the 1990's). Due to the large negative value for Australia, a net forest area loss of -243,000 ha is reported for the subregion during the 1990's.

Other Oceania. Fiji, Samoa and Papua New Guinea show a small negative difference between AreaChange_FRA2000 and AreaChange_FRA1990. Vanuatu, Solomon Islands and New Caledonia show positive changes. The figures for the remaining countries are marginal. The net result for the subregion is a small positive difference between AreaChange_{FRA2000 and} AreaChange_FRA1990; meaning continued forest area loss during the 1990's but at a slightly reduced rate compared to the 1980's.

Europe

Positive difference for all four subregions (Northern, Central, Southern Europe and Belarus, Moldova, Other Former USSR & Ukraine).

Northern Europe. The differences between AreaChange_{FRA2000 and} AreaChange_FRA1990 for Iceland, Finland, Norway and Sweden are all small and positive. The trend may be due to a combination of regrowth of marginalized agricultural land /pastures by natural regrowth and by plantation.

Central Europe. Austria, Germany, Hungary, Switzerland and United Kingdom reported negative differences between AreaChange_FRA2000 and AreaChange_FRA1990. For Germany the FRA data indicates an increasing forest area in the 1980's followed by a constant forest area through the 1990's. For Austria, Hungary, Switzerland and United Kingdom the FRA data indicate a forest area that continued increasing through the 1990's but at a slower rate than in the 1980's.

France, Former Czechoslovakia, Ireland, and Poland show a faster rate of increase of forest area during the 1990's than during the 1980's.

For Denmark and Netherlands there was no difference in the reported rate of forest area change from FRA 1990 to FRA 2000.

For Belgium /Luxembourg a small reported annual increase in forest area reported for the 1980's was followed by a small annual decrease in forest area during the 1990's.

Southern Europe. All countries, except Albania and Former Yugoslavia SFR report a positive difference between AreaChange_FRA2000 and AreaChange_FRA1990. Albania reverted from a small reported annual increase during the 1980's to an annual decrease of app. 8,000 ha during the 1990's, assumed primarily to be due to conversion to agriculture along with improved national data and changed national definitions. For Former Yugoslavia SFR an annual increase during the 1980's of 34,500 ha was followed by a virtual stagnation during the 1990's (AreaChange_FRA2000 reported at 3,000). The difference in annual forest area change of -31,500 ha reported from the 1980s to the 1990s may be a result of civil war and the subsequent division into several national units for FRA 2000.

Bulgaria, Greece, Italy, Portugal, Romania and Spain all reported small positive figures for AreaChange_FRA1990, and a larger positive figure for AreaChange_FRA2000. The biggest positive difference from the 1980's to the 1990's for the countries of the subregion is for Spain, where AreaChange_FRA2000 was reported as 85,100 ha. Change of FRA definitions between the assessments may in part explain the pattern.

Belarus, Ukraine and Other Former USSR. All reporting Units are characterized by a positive difference between AreaChange_FRA2000 and AreaChange_FRA1990. This is especially the case for Other Former USSR where the difference is reported at 453,000 ha. The vast areas, the political and economical changes and the changes in FRA definitions along with the fact that AreaChange_FRA1990 was interpreted as 0 for the present comparison are factors that may contribute in explaining the difference.

For Belarus the difference is also large and positive (228,700 ha), much for the same reasons. For Belarus however AreaChange_FRA1990 was reported at 27,300 ha. The difference between AreaChange_FRA2000 and AreaChange_FRA1990 for Ukraine amounts to 7,000 ha.

North & Central America

Positive differences for all 3 subregions (North America excl. Mexico, Central America & Mexico and Caribbean).

North America excl. Mexico. Canada reported no difference between AreaChange_FRA2000 and AreaChange_FRA1990 as both figures were reported as n.s. The figures for USA indicated a large annual decrease in forest area during the 1980's followed by a reported large annual increase during the 1990's (AreaChange_FRA1990 reported at -316,500 ha and AreaChange_FRA2000 reported at 388,000 ha).

Central America and Mexico. AreaChange_FRA2000 is negative for all countries of the region. The same applies for AreaChange_FRA1990. For Belize and El Salvador AreaChange_FRA2000 is smaller than AreaChange_FRA1990 indicating a reported acceleration in the rate of forest area loss from the 1980's to the 1990's which may be due to conversion to other land use. For Costa Rica, Guatemala, Honduras, Mexico, Nicaragua and Panama there is a positive difference between AreaChange_FRA2000and AreaChange_FRA1990 indicating a decrease in the rate of reported forest area loss from the 1980's to the 1990's. This trend may primarily be due to improved information on the forest resources becoming available for FRA 2000 along with the change of methodology for assessing the change.

Caribbean. In global context the forest area of the Caribbean is small. 14 countries report AreaChange_FRA2000 as being non-significant (for twelve of these AreaChange_FRA1990 was also reported as 0 ha). Bahamas and Dominican Republic showed a negative development of forest area in the 1980's followed by stability during the 1990's. Cuba which has the largest forest area in the subregion reported a positive development of annual forest area change between the two assessments (+31,500 ha) resulting from a negative AreaChange_FRA1990 followed by a positive AreaChange_FRA2000. For Jamaica and Trinidad & Tobago AreaChange_FRA2000 was negative but the rate of deforestation had decreased from the 1980's. AreaChange_FRA1990 for Guadeloupe was reported at 0 ha while for AreaChange_FRA2000 2,000 ha was reported, i.e. a small net increase.

AreaChange_FRA1990 for St. Lucia was reported at 0 ha while for AreaChange_FRA2000 -1,000 ha was reported i.e. a small net decrease. Haiti reported negative forest area changes during the 1980's as well as during the 1990's with the rate of forest area loss increasing from the 1980's to the 1990's.

South America

Large positive difference reported for Tropical South America. Small negative difference reported for Non Tropical South America.

Tropical South America. AreaChange_FRA1990 and AreaChange_FRA2000 for French Guiana were reported as 0 ha.

For Bolivia, Brazil, Colombia, Equador, Paraguay, Suriname and Venezuela AreaChange_FRA1990 and AreaChange_FRA2000 were both negative, but the reported rate of deforestation had decreased from the 1980's to the 1990's. With a difference of 1,166,600 ha, Brazil is the most significant contributor to this reported change. Other main contributors are Bolivia (463,000 ha), Venezuela (364,400 ha) and Paraguay (279,300 ha).

For Guyana and Peru the difference between AreaChange_FRA2000 and AreaChange_FRA1990 was negative indicating an increased annual rate of net forest area loss from the 1980's to the 1990's.

The summed figures for Tropical South America indicate a total positive difference between AreaChange_FRA2000 and AreaChange_FRA1990 of about 2,5 million ha, i.e. a reported large decrease in the rate of forest area loss from the 1980's to the 1990's, mainly due to the above figures for Brazil, Bolivia, Venezuela and Paraguay. The apparent improvement may in part be explained by improved information becoming available for most countries between FRA 1990 and FRA 2000 along with the change of methodology between the assessments for assessing the forest area change.

Non Tropical South America. Argentina and Chile reported negatives differences between AreaChange_FRA2000 and AreaChange_FRA1990, meaning an increase in the rate of net forest area loss from the 1980's to the 1990's. Uruguay reported a relatively large annual increase (50,000 ha) in forest area during the 1990's.

Main contributors to difference in Forest Area Change between FRA 1990 and FRA 2000

The 26 countries in table 3-02 exhibit differences (positive as well as negative) between AreaChange_FRA2000 and AreaChange_FRA1990 that exceed 100.000 ha. Positive and negative values occur both in industrialized and developing countries. Together they account for a net difference of annual forest area change between the 1980's and the 1990's (according to FRA 1990 and FRA 2000 respectively) of 3.2 million ha (or 85 % of the total difference of the net annual forest area change between the assessment [3.7 million ha]).

Table 3-02. Annual forest area change for countries where differences between AreaChange_FRA2000 and AreaChange_FRA1990 exceed 100.000 ha.

Country	FRA 1990 Figures				FRA 2000 Figures
	Total forest 1980	Total forest 1990	Forest area change 1980-1990		Total forest 1990	Total forest 2000	Forest area change 1990-2000		Difference in annual change
			Annual change				Annual change
	000 ha	000 ha	000 ha	%	000 ha	000 ha	000 ha	%	000 ha
Cameroon	21,574	20,366	-121	-1	26,076	23,858	-222	-1	-101
Dem. Rep. Congo	120,611	113,317	-729	-1	140,531	135,207	-532	0	197
Côte d'Ivoire	12,125	10,967	-116	-1	9,766	7,117	-265	-3	-149
Gabon	19,408	18,256	-115	-1	21,927	21,826	-10	0	105
Nigeria	16,938	15,785	-115	-1	17,501	13,517	-398	-3	-283
Sudan	47,911	43,179	-473	-1	71,216	61,627	-959	-1	-486
United Republic of Tanzania	38,003	33,709	-429	-1	39,724	38,811	-91	0	338
Zambia	35,958	32,349	-361	-1	39,755	31,246	-851	-2	-490
Zimbabwe	9,577	8,981	-60	-1	22,239	19,040	-320	-2	-260
India	58,259	64,959	670	1	63,732	64,113	38	0	-632
China	126,401	133,799	740	1	145,417	163,480	1,806	1	1,066
Indonesia *	124,476	115,674	-880	-1	118,651	105,493	-1,315	-1	-435
Malaysia	21,561	17,664	-390	-2	21,661	19,292	-237	-1	153
Myanmar	32,905	29,091	-381	-1	39,588	34,419	-517	-1	-136
Philippines	11,194	8,034	-316	-4	6,676	5,789	-89	-1	227
Thailand	18,120	13,264	-486	-4	15,886	14,762	-112	-1	374
Vietnam	10,662	9,782	-88	-1	9,303	9,819	52	1	140
Australia	39,831	39,837	1	0	157,359	154,539	-282	0	-283
Belarus	5,743	6,016	27	0	6,840	9,402	256	3	229
Other Former USSR *	739,729	739,729	0	0	877,568	882,077	452	0	452
USA	212,738	209,573	-317	0	222,113	225,993	388	0	705
Bolivia	55,585	49,345	-624	-1	54,679	53,068	-161	0	463
Brazil	600,763	566,007	-3,476	-1	566,998	543,905	-2,309	0	1,167
Colombia	57,771	54,190	-358	-1	51,506	49,601	-190	0	168
Paraguay	16,891	12,868	-402	-3	24,602	23,372	-123	-1	279
Venezuela	51,767	45,943	-582	-1	51,681	49,506	-218	0	364
Total	2,506,501	2,412,684	-9,382		2,822,995	2,760,879	-6,209		3,173

* Defined as mentioned in table 2-01

3.3 Summing up the comparison of forest area change

The net difference between the net annual Forest Area Change estimates of FRA 1990 (- 13.1 million ha) and FRA 2000 (- 9.4 million ha) is 3.7 million hectares.

As opposed to the difference between Area90_FRA2000 and Area90_FRA2000, which owed to large differences in the reported forest areas for a few countries, the picture for forest area change is more complex. This net difference of 3.7 million ha is made up of a number of country contributions that assume both positive as well as negative values.

On a regional level the data for Africa and Oceania show a negative development in the net forest area change rates from the 1980's to the 1990's, meaning an apparent acceleration of the deforestation rate in these regions from the 1980's to the 1990's.

In South America, North and Central America and Asia the net rate of deforestation as reported to FRA appears to have slowed down from the 1980's to the 1990's at regional level (most explicit for South America with a positive difference of about 2,5 million ha.).

In the case of Europe (including former USSR) the reported rate of forest area increase has increased from the 1980's to the 1990's.

Behind the regional figures are a number of subregional and national trends that are often in opposite directions.

A direct comparison between the figures for forest area change from FRA 1990 and FRA 2000 is complicated by the following differences between the assessments:

_ Change in the methodology for generating Forest Area Change rates for developing countries. FRA 1990 used a deforestation model for generating forest area change estimates. FRA 2000 used country data, where possible from two or more points in time. The use of the deforestation model was discontinued for FRA 2000 as it was acknowledged that the correlation between forest area change and the demographic parameters used as model input is weak at national level. The use of the model for the FRA 1990 estimates of forest area change may, seen in retrospect, have overestimated deforestation figures for some tropical countries for the 1980's. This change of methodology may in part explain the seemingly large positive development for South America.

_ Lack of reliable data for AreaChange_FRA2000. For less than half of all countries, time series with high comparability between the observations have been used for estimating AreaChange_FRA2000. For many countries FRA 2000 had to rely on secondary source information and/or expert opinion for producing the estimates. Comparison with FRA 2000 remote sensing survey indicates that especially for tropical Africa the lack of reliable baseline data seems to lead to an overestimation of the rate of deforestation reported for FRA 2000, which may in part explain some of the negative trend for Africa.

_ Availability of improved data. In spite of the above mentioned shortcomings in the source data for the FRA 2000 forest area change estimates, the underlying source data and consequently the FRA 2000 estimates, still represents an improvement from FRA 1990. This is especially the case for countries where AreaChange_FRA1990 is reported as not available, e.g. Other Former USSR. Improved information becoming available between FRA 1990 and FRA 2000 resulted in drastic revisions of the figures for forest area change between FRA 1990 and FRA 2000, e.g. Australia and United Republic of Tanzania. This availability of improved data for FRA 2000 may explain part of the positive development observed in Europe (mainly due to large positive figures reported for Other Former USSR) and may partly explain the negative trend for Oceania (mainly due to large negative figures reported for Australia).

_ Change of FRA definitions from FRA 1990 to FRA 2000. FRA 2000 was the first assessment to employ a homogenous set of definition of forest globally. The change of threshold value for forest area of industrialized countries from 20 % used for FRA 1990 to 10 % used for FRA 2000, had a major impact on the Forest Areas of certain countries e.g. Australia and Other Former USSR. However the effect of changed definitions on forest area change rates seems to be less, likely because the forest area change rates of FRA 1990 and FRA 2000 are based on data from the respective assessment only, thereby eliminating differences owing to using two datasets based on different definitions for estimating forest area change.

Chapter 3 presented and compared the reported figures for forest area change based on country submissions from FRA 1990 and FRA 2000 at regional and subregional level. The main trends concerning reported differences in the figures have been described and the main contributors identified. Forest area change data is available for the individual country in appendix II.

The above mentioned differences between the assessments complicate a straightforward comparison of the forest area change figures from FRA 1990 and FRA 2000. These are analysed further in chapter 4 along with adjustments for AreaChange_FRA2000 for Africa and adjustments for an estimated 70% success rate of plantations.

⁴ For the industrialized countries FRA 1990 reports annual changes for forest and other wooded land as one figure. For industrialized countries FRA 1990 defined forest as areas with a crown cover of >20%. FRA 2000 employs a common global set of definitions defining forest as areas with crown cover >10% and with a woody vegetation of minimum 5 meters height (for FRA 1990 crown density and height criteria was 20% and 7 meters respectively). The two definitions are not fully comparable, but a comparison is still attempted as part of the change of the FRA 1990 other wooded land category will be in the interval 10-20% crown density and 5-7 meters height and thus technically be defined as forest according to FRA 2000 definitions.

⁵ Increase, meaning assuming a greater positive value or a smaller negative value or changing from a negative to a positive value between FRA 1990 and FRA 2000.

⁶ Decrease, meaning assuming a greater negative value or a smaller positive value or changing from a positive to a negative value between FRA 1990 and FRA 2000.

⁷ net annual forest cover change refers to the fact that the figures encompass positive and negative changes in both natural forest area and in forest plantation area (including rubber for the FRA 2000).

⁸ These apparent differences can however not be confirmed by FRA 2000 Pan Tropical Remote Sensing Survey, which was carried out independently of the country data compilation. The Remote Sensing Survey did not find any statistically significant differences between net forest cover change rates between the 1980's and the 1990's for any of the tropical regions.