Previous Page Table of Contents Next Page


6. BIOMASS ESTIMATES FROM GIS MODELING


6.1 GENERAL APPROACH
6.2 EXAMPLES OF RESULTS

A new approach for estimating biomass density will be presented in this section that is based on a modelling method with GIS (geographic information systems) technology. It uses various existing digital data bases and maps of reliable inventories, population density, climate, vegetation, ecofloristic zones, soils, and topography. This method was developed as a means to extrapolate reliable inventory data that is generally limited in area coverage to biomass density estimates at larger scales, such as continents.

6.1 GENERAL APPROACH

Forest biomass density has been modelled in a multi-stage approach using GIS software packages and a variety of spatial and statistical data bases. For estimating forest biomass density using a GIS, the present distribution of forest biomass density was assumed to be based on the potential amount that the landscape can support under prevailing environmental conditions, and the cumulative impact of human activities on forests that reduce its biomass density. Many spatial data layers have been developed from existing data bases or were prepared by specialists (e.g., for the FRA 1990 Project; FAO 1993). These data layers were entered into a GIS (in this case ARC/INFO-GRID) and processed according to specifications of the model. The following GIS data layers, or digital maps, and statistical data bases were used as inputs to the modelling effort (further details are given in Brown et al. 1993, Iverson et al. 1994):

· Climatic index map based on FAO meteorological station data
· Precipitation map
· FAO soils map, recoded into depth, textural, and slope classes
· Topography map
· FRA 1990 Ecofloristic Zone maps
· FRA 1990 vegetation maps
· National and subnational boundary map
· Subnational population data (1970, 1980, and 1990) from FAO
· Biomass density estimates from reliable forest inventory data using the above methods

The first step in this analysis was to estimate a potential biomass density (PBD) for forests. This was accomplished by first developing an index of potential biomass density based on climatic, edaphic, and topographic factors. The potential biomass density map was masked with a forest map, produced by reclassifying all the forest classes of the FRA 1990 vegetation maps into one forest class. The potential biomass density index (PBI) was calculated according to a simple model, based on overlaying the following GIS data layers:

PBI = climatic index + precipitation + soil (texture, depth, slope) + topography

Each of these factors was spatially represented by a numerical scale whose values were ranked according to how the particular factor affected forest biomass (details of the scaling are given in Iverson et al. 1994). The digital maps were overlain according to the above model and the results calibrated and validated using existing forest inventories for mature forests, literature sources for small scale ecological studies, and the FRA 1990 ecofloristic zone (EFZ) map (Figure 2).

Figure 2 - Map of estimates of potential biomass density classes for forests of tropical African countries. The units are in Mg/ha which is the same as t/ha (1 Mg = 106 g = 1 ton).

The final step was to add the influence of all human activities that result in a reduction of biomass of forests. This step was accomplished by using the biomass estimates from the reliable forest inventory data. The first step was to calculate degradation ratios, defined here as biomass density estimates from the inventories (representing all forests of a subnational unit) divided by the potential biomass density for all forests in the same subnational unit. Several models were tried, but the best results (highest r2) were obtained from linear models of degradation ratios versus the natural logarithm of population density of the subnational unit, stratified into main ecoregions or forest types (e.g. closed forests and woodlands). A similar relationship was found for forest cover and population density stratified by ecological and geographic zones (FAO 1993). The significant relationships between degradation ratio or forest cover and population density are not meant to imply that the relationships are causal; many other socio-economic factors are involved. However, the empirical relationships are useful because data on population density are generally readily available, but reliable data on other socio-economic factors are often not so available.

Figure 3 - Relationship between degradation ratios (biomass density estimates from the inventories divided by the potential biomass density for all forests in the same subnational unit) and the natural logarithm (Ln) of the population density (number of persons/km2) for (a) closed forests and (b) open forests and woodlands. The second row of numbers on the x-axis of (a) are the actual population densities corresponding to the natural logarithm values.

(a) Closed forest

(b) Woodland

The data and regression equations used for the map of Africa are shown in Figure 3. The data base for these regression equations is a combination of data from inventories in tropical Asia and Africa. The estimates of degradation ratios for a given population density for these two regions overlap well perhaps implying a similar pattern of forest use. These regression equations were used with the potential biomass density map and population data to produce a map of degradation ratios by subnational regions for each tropical African country as shown in Figure 4. The degradation ratio map was then used with the potential biomass density map to produce a map of actual forest biomass density. As with all regression equations, the ones shown in Figure 3 should not be extrapolated too much beyond the data. Degradation of forest biomass density is not expected to continue according to the trends shown in Figure 3 at high population densities. For example, in the more industrialized countries, deforestation and degradation have been reversed irrespective of population development as other socioeconomic factors exert more influence.

6.2 EXAMPLES OF RESULTS

To date, this method has been completely applied to the tropical Asian and Africa regions only. An example of the map of actual biomass density for tropical Africa is shown in Figure 5. Verification of this approach with inventory data not used in the model suggests that it is sound. However, the results are presently limited by the quality of the input data such as the small number and coverage of reliable forest inventories, FAO soils map, and the current vegetation maps. Improvements in reliability of estimates will be brought about with improvements in these input data.

The results of the GIS modelling can be by region, subregion, or country. They will usually consist of area-weighted biomass density estimates (t/ha) summarized by various vegetation and ecofloristic zones (see below). Because of the low resolution in the input data bases, reporting below the national scale will result in decreasing reliability of results. It is therefore justified only in large countries such as India or Zaire.

Estimates of potential and actual biomass (density and total), as determined by modelling in a GIS, for most of the tropical countries in tropical Asia and Africa are shown in Table 6.1. Many of these values can be compared to the biomass density estimates given in Section 5 as a means of validating the modelling results.

Figure 4 - Map of degradation ratios, grouped into five classes, for subnational units of tropical African countries which contain forests.

Figure 5 - Map of estimates of actual biomass density classes for forests of tropical African countries. The units are in Mg/ha which is the same as t/ha (1 Mg = 106 g = 1 ton).

As of about 1980, the actual biomass densities of forests in tropical Asia were about 50% of their potential densities. Forests of Brunei, Cambodia, and Sarawak/Sabah have the highest actual biomass densities, >300 t/ha, representing about 58-72% of their potential amount. At the other end of the scale are the forests of Bangladesh and India that have actual biomass densities of<170 t/ha, or about 37% of their potential. Forests in these two countries appear to have been subject to high human pressure that has reduced the forest stock to very low values on average.

Of the 37 African countries with tropical forests, about half contained forests that appear to contain less than 60% of their potential biomass density. On average, forests in Congo, Equatorial Guinea, Gabon, and Liberia have the highest actual biomass densities at > 300 t/ha which represents about 65-92% of their potential biomass density. All these countries are dominated by lowland moist forests. Countries dominated by very dry climates such as Botswana, Somalia, Zimbabwe, and countries of the Sahel have the lowest actual biomass densities, <50 t/ha.

Table 6.1

Area-weighted average potential (without human impacts) and actual (with human impacts as of about 1980) biomass density estimates (t/ha) and degradation ratios (DR = inventory biomass density divided by potential biomass density) for tropical Asian and African countries.

Country/Region

Potential (t/ha)

Actual (t/ha)

DR

ASIA

Bangladesh

463

170

0.37

Brunei

577

382

0.66

Cambodia

419

301

0.72

India

348

129

0.37

Indonesia

533

262

0.49

Laos

342

272

0.80

Myanmar

388

231

0.60

Malaysia





Peninsular

518

210

0.41


Sarawak/Sabah

571

331

0.58

Philippines

511

223

0.44

Sri Lanka

413

200

0.48

Thailand

356

185

0.52

Vietnam

372

262

0.70

Average (area weighted)

437

224

0.5?

AFRICA

Angola

100

73

0.73

Benin

112

58

0.52

Botswana

15

13

0.91

Burkina Faso

65

34

0.53

Burundi

119

43

0.36

Cameroon

307

217

0.71

CAR

243

200

0.68

Chad

63

43

0.68

Congo

374

344

0.92

Cote d'Ivoire

276

165

0.60

Equatorial Guinea

442

318

0.72

Ethiopia

101

52

0.51

Gabon

375

339

0.90

Gambia

64

29

0.45

Ghana

182

83

0.45

Guinea

259

140

0.54

Guinea Bissau

153

85

0.55

Kenya

58

33

0.57

Liberia

466

305

0.65

Madagascar

322

196

0.61

Malawi

108

47

0.44

Mali

75

45

0.60

Mozambique

96

57

0.60

Niger

16

9

0.53

Nigeria

128

49

0.38

Rwanda

103

34

0.33

Senegal

50

32

0,62

Sierra Leone

411

199

0.48

Somalia

20

13

0.63

Sudan

95

64

0.67

Tanzania

83

45

0.55

Togo

155

72

0.46

Uganda

237

102

0.43

Zaire

297

206

0.69

Zambia

67

47

0.70

Zimbabwe

26

14

0.51

The biomass data were also summarized by ecoregions (reclassified ecofloristic zone map) as shown in Table 6.2. The general trends in biomass density by ecoregion for tropical Asia and Africa are consistent with expected patterns of biomass distribution: decreasing biomass density with decreasing moisture and increasing elevation. Most of the total biomass in continental Asia is in the lowland seasonal forests (52%) and only a small fraction is in the lowland dry forests (2%). However, in insular Asia, 89% of the total biomass is in the lowland moist ecoregion with only a trace in the seasonal zone. Forests in the continental dry zone and the insular seasonal zone appear to be the most degraded with actual biomass densities that are about 30-40% of their potential.

In general, forests in a given ecoregion of Africa have degradation ratios closer to 1.00, i.e., forest are less degraded than those of Asia (Table 6.2), presumably because of the lower population densities in the forested regions of Africa at present. The degradation ratios indicate that degradation increases with increasing aridity in both lowland and montane ecoregions of Africa, and that montane zones are more seriously degraded than lowland zones. The more advanced degradation of the montane zones of Africa probably reflect the more favorable climate that is generally preferred for human habitation and agriculture.

Table 6.2

Area-weighted average potential (without human impacts) and actual (with human impacts as of about 1980) biomass density (t/ha), and degradation ratios (DR = inventory biomass density divided by potential biomass density) for forests of tropical Asia and Africa by ecoregion.

Ecoregion

Potential

Actual (t/ha)

DR

CONTINENTAL TROPICAL ASIA:

Lowland moist

449

225

0.50

Lowland seasonal

350

187

0.53

Lowland dry

244

76

0.31

Montane moist

353

222

0.63

Montane seasonal

306

155

0.51

INSULAR TROPICAL ASIA

Lowland moist

543

273

0.50

Lowland seasonal

452

174

0.38

Montane moist

504

254

0.50

TROPICAL AFRICA

Lowland moist

412

299

0.73

Lowland seasonal

211

141

0.67

Lowland dry

92

60

0.65

Lowland very dry

33

20

0.61

Montane moist

197

105

0.53

Montane seasonal

78

37

0.47


Previous Page Top of Page Next Page