Lengths of retrospective periods of documented LULCC, as well as reliability of data, vary by continents and countries. Relatively reliable 'reconstructions' of global land-use dynamics have been done since the 1850s. According to these estimates, between 1850 and 1990, the area of cultivated lands, worldwide, has been estimated to have increased by more than a factor of 4, from 320 x 106ha in 1850 to 1360 x 106 in 1990. The most rapid increase occurred in tropical regions after the 1940s - about half of the increase of ^1000 x 106ha occurred during this period. About 730 x 106ha of agricultural lands was cleared from forests and woodlands that reduced the area of the world forests by 17% since 1850. The total net flux of carbon to the atmosphere from changes in land use was estimated to be 124 Pg C over the period 1850-1990. Changes in the forest area accounted for almost 90% of the net long-term carbon flux.
Currently, there are two major sources for monitoring LULCC including deforestation and degradation: RS and data of national forest inventories. RS was provided either at global or biome scale using imagery of coarse spatial resolution, or by statistical sampling of fine resolution. The basic tradeoff between these two groups of RS instruments is between spatial and temporal resolution. For
Table 1 Annual average rates of tropical deforestation (106hayr~1)
Average annual rates of deforestation Region 1980sa 1990sa example, Landsat's revisit time is 16 days - for satellites of coarse resolution, near-daily. Due to the high probability of cloud cover in many regions and the presence of smoke from vegetation fires, instruments of fine resolution cannot provide a satisfactory complete coverage. However, fine-resolution sensors cannot be avoided due to patchy structure of tropical deforestation with many small plots. Currently, a number of different satellites (optical bands) are used: SPOT (20 m resolution), IRS-2 (6-56 m), Landsat 5 and 7 (30 m), Terra (250-1000 m), ENVISAT (300 m), some others. Use of radars, which can penetrate the cloud cover, is one of the alternatives for estimating deforestation by satellites. Applications of radars from new satellites (e.g., SAR from ENVISAT, 75 m resolution, or ALOS, 50 m) show promising results. However, identification of small patches of deforested land, distinguishing degraded forests, and indicating regrowth can still not reliably be done from space for large areas. Thus, available estimates of global deforestation are not consistent and reliable enough yet.
Reported areas of deforestation in the Tropics vary substantially (Table 1). Several subsequent estimates of dynamics of the global forest cover were provided by FAO FRA, mostly using country surveys which are based on compilation and standardization of data of national statistics. Recently, FRA-2005 presented revised estimates for 1990-2000 and new results for 2000-05. As a total conclusion for 1990-05, the global deforestation rate was estimated at 13 x 10 ha per year, almost completely in the Tropics. Taking into account increased areas of forest plantations and natural expansion of forests, particularly in temperate and boreal zones, the global net change of forest area was estimated at -8.9 x 106ha in 1990-2000 (equivalent to loss of 0.22% to remaining area annually) and -7.3 x 106ha in 2000-05 (-0.18%). The largest net loss of forests in 2000-05 was estimated at 4.6 x 10 hayr-1 for Central and Southern America followed by Africa, which lost 4.0 x 106 ha annually. During the last 5 years, the previously negative trend in Asia has reversed due to large-scale plantations established mostly
Net loss of forest area 2000-05b 1990-2000 2000-05c
aRange due to available publications and results of surveys.
bFAO (2005) data estimated as the total area for countries with net loss of forest area; data for Asia additionally include 0.4 x 106 ha for Oceania. cFAO (2005) Global Resources Assessment 2005. Progress towards sustainable forest management. FAO Forestry 147, 350pp. Rome: Food and Agriculture Organization of the United Nations; the area of net change of forest area for USA and Canada are deducted from the total area for the American continent.
in China and India. Other estimates of tropical deforestation for the last two decades of the twenty-first century vary from 8 x 106 to >16 x 106hayr-1. RS estimates report that over 20 years (from the 1970s to the 1990s) the area of global forest decreased by 6%. On average, the RS estimates report lesser areas of tropical deforestation than FAO estimates. Likely, the above estimates of deforestation rate are slightly overestimated due to the fact that national inventories and RS data do not adequately record the regrowth. However, from another side, small deforested patches and selective logging, as a rule, are not included in the reported area. Probably, an aggregated conclusion on the current level of deforestation in the Tropics of c. 10 x 106hayr-1 can be considered as 'the best' conservative estimate of this process.
Considering the regional aspect, Brazil reported 21% of the net global loss for 1990-2000 and 24% for 2000-05, but this country has probably the best national RS system of deforestation monitoring: since 1997, the Brazilian National Institute of Space Research (INPE) has been monitoring deforestation down to 6.25 ha. Estimated areas for the three years 2002-05 (August to August) were on average 2.37 x 106hayr-1 with reported error ±4%. Overall, during the last 25 years, the Brazilian Amazon lost an area of forest greater than the size of Germany. For ten countries of Southeast Asia, about 2.3 x 106ha of forests was cleared every year between 1990 and 2000 and transferred to other forms of land use. Annual deforestation in Indonesia was estimated some 1.7 x 106ha in 1987-97 with the increase to 2.1 x 106 ha in 2003.
Was this article helpful?