Prospects for intensifying land use while protecting forests
“Best bet” Land-use Systems
Country reports
Alternatives to Slash-and-Burn in Brazil
Trends In Land Use Patterns And Impacts
Unique id: IDA0QFXB
Source file: D:\Projects\ASB\ASB Country and Thematic reports\Brazil country report\ASB Brazil Summary Report.xml
Authors: S. Vosti, C. L. Carpentier, J. Witcover, . Carvalho dos Santos, E. Muñoz Braz, J. Ferreira Valentim, S. J. de Magalhães de Oliveira, C. Palm, F. de Souza Moreira, A. Cattaneo, A. Gillison, A. Mansur Mendes, V. Rodrigues, T. C. de Araújo Gomes, M. V. Neves d’Oliveira, E. do Amaral, S. Fujisaka, C. Castilla, T. Tomich, D. Bignell, D. Gonçalves Cordeiro, A. Hermes Vieira, R.S. Correira da Costa, M. Faminow, M. Locatelli, M. Swift, S. Weise, M. van Noordwijk, N. Sampaio, I. L. Franke, H. J. Borges de Araujo, L. M. Rossi, E. Barros, B. Feigl, S.P. Huang, J. Cares, C. Pinho de Sá, . Carneiro, P. Woomer
--------------------------------------------------------------------------
The underlying question that motivates our examination of trends in deforestation and land use is whether farmers can derive better livelihoods from their land when this is kept as forest. Earlier sections highlighted farmers’ objectives in adopting a particular LUS on a particular plot of land. Such decisions, made repeatedly on all farmers’ plots, are a major factor determining the speed at which forest falls and the costs and benefits of forest conversion. This subsection explores this dynamic in more detail.
Smallholder land use
patterns
Figure 5 (in Section 1.6) sets out the land use trajectory of a plot of land from forest to its end use as pasture, showing the amounts of time that different uses typically remain in place. The figure notes the observed periodicity of forest felling and the average size of plot felled on sample farms in the project areas during the 1994 field survey. As noted in Section 1.6, smallholders on average deforest about 4.7 ha of forest every other year. Private lots are usually deforested from the front of the lot (facing the road) to the back (Fujisaka et al, 1996), with the area in pasture steadily accumulating.
Figures 14 and 15 documents the conversion from forest to pasture in smallholders’ lots, based on recall data from the 1994 and 1996 surveys. On average, holdings that were 88% forested upon their owners’ arrival were only 61% forested in 1994 (Figure 14) and 56% forested in 1996.[1] The 1996 average includes the 35% of the sample farms that had less than half their operational holding still in forest and the 10% that had less than a quarter still forested. While 60% of sample farmers reported deforesting every second year and an additional 25% reported deforesting every third year, the patterns revealed by data analysis often deviated from reported frequencies. Deforestation was less prevalent in 1996 (when about a third of the sample cut down some forest and nearly a quarter felled secondary forest fallow) compared with 1994 (when 60% of the sample deforested and nearly 70% razed secondary forest). Thus, the spike in deforestation rates observed by some researchers in the Amazon in 1994 and 1995 (Lele et al, 2000; INPE, 2000) also emerged in this sample. The mean area felled for all farmers dropped from 2.5 ha in 1994 to 1.5 ha in 1996 for the forest, and from 3 to 1 ha over the same period for secondary forest re-growth. Among those who deforested, however, the mean area felled held steady at between 4 and 4.5 ha in both years for both forest and re-growth, although with substantial variation across households. Over the entire period since their arrival, 1996 owners had deforested on average 3.1 ha per year, but in many cases lots had been deforested at an average rate of 2.5 ha per year since their initial settlement by previous owners. Lots opened more recently, but which had not yet changed hands, had significantly higher average annual deforestation rates.
Figure 14 Land uses on sample farms, 1994
Source: ASB field data, 1994
Figure 15 Changes in land use on sample farms, 1994 to 1996
Source: ASB field data, 1994-97
Pasture, on the other hand, grew from an average of 3% of the land, when settlers arrived, to 21% in 1994 (Figure 14) and 27% in 1996. Between 1994 and 1996 alone (Figure 15), forest area decreased by a mean of 7% of the operational holding at the household level, while pasture area rose by 10%. There was also a slight expansion in the area of perennial systems.
A snapshot of land use (taken in 1996) on farms initially settled at different times (of different ‘vintages’, in Figure 16) confirms a striking pattern in line with expectations if the land use trajectory described above, from forest to pasture, were played out continuously on one plot of land after another. ‘Old’ farms have much less forest than ‘young’ farms; and while deforestation patterns are less obvious for some ‘middle-aged’ farms, a clear and positive link between ‘time since opening’ and area in pasture is evident for all farms.
Figure 16. Land use by farm ‘vintages’
Cross-sectional field data show that sample farmers on average generate outputs of higher value from cleared than from forested land in a given year. This too is expected, given the land use trajectories described above. Using field data on the land uses adopted by an average farm household and market prices for 1994 as a basis, the value of total output (VTO) for the 1993/94 season was estimated to have averaged R$ 3448,[2] which is a reasonable proxy for net returns given the low level of production costs. Figure 17 reports the distribution of on-farm VTO across groups of activities: just under half (45%) of VTO was derived from cattle-based activities (milk plus the value of growth in animals younger than five years), while 47% was from annual cropping and less than 10% from extractive activities. Analysis for 1996 yielded similar results.
Figure 17. Distribution of on-farm income by activity
For the average family size in the sample (five members), a return of R$ 690 per capita (R$ 3448 divided by 5) ranks above the Brazilian minimum wage as well as the World Bank estimated poverty line for 1995, indicating that, on average, farmers in projects roughly 15 to 25 years after their establishment have little incentive to leave farming to enter the off-farm labour force (Vosti et al, 2002; Faminow et al, 1999).
In summary, the analyses of field data suggest that, in the absence of major changes in the prices, policies, institutions and technologies prevailing in the region, the area in pasture will continue to increase, while that in forest will continue to decline and swidden agriculture, with a long fallow period, will not be practised (Vosti and Valentim, 1998). A potential candidate land use with promise to slow deforestation—perennial cropping—may have appeared on the horizon, but more area in perennial crops does not necessarily guarantee more area in forest. Rather, the decision to invest in pasture or perennials appears finely balanced, with the balance at present tipping towards pasture. Evidence suggests that, in the years leading up to 1996 at least, a mixed pasture/perennial farm had a significantly lower standard of living than the more prevalent pasture-dominated farm. Hence, the area in perennials is not expected to increase significantly as a proportion of cleared area in the foreseeable future.
The general trend in land use—of conversion from forest to
pasture—held despite evidence of substantial variation in production
technologies, particularly in the case of pasture. There
was some evidence of intensification via intercropping and improved pasture
management, while intensification via purchased inputs was rare.
Deforestation rates varied but seemed to have accelerated since time from opening, with more than a few farms crossing the ‘50%-of-farm-in-forest’ barrier decreed by law (in effect at the time of the survey, but rarely enforced). This finding is in keeping with the idea that pressures to deforest are greater in an environment with higher populations and access to markets than in one where farmers must rely on their land and household labour alone for subsistence. However, while the amount of land cleared on each occasion stayed relatively constant, the timing of clearing varied—providing a potential entry point for policy measures seeking to slow deforestation by reducing the frequency of clearing.
Some farmers in the sample were quite well-off—buying new lots and consumer durables, achieving yields comparable to those of research stations and selling their output. Other farmers had fewer signs of wealth, lower yields and more limited access to, and participation in, labour and output markets. This suggests an important bimodality in the sample in terms of those who are succeeding and those who are not. While the gap between successful and unsuccessful farm households is expected to persist, the proportion of less successful farmers in this rural population is expected to decline, since members of this group will tend to migrate to other areas, rural and urban.
Future trends
suggested by the ASB matrix
The results for