Results And Discussion

The effects of the conversion of tropical forest to pasture on total soil C, using current weather data (Table 9.1), was analyzed in detail by Cerri et al. (2003), using the Century model and chrono-sequence data collected from the Nova Vida Ranch. First, the model was applied to estimate equilibrium organic matter levels, plant productivity, and residual C

inputs under native forest conditions. Then Century was set to simulate the deforestation following slash and burn. SOM dynamics were simulated for pastures established in 1911, 1951, 1972, 1979, 1983, 1987, and 1989.

Using input data from Nova Vida Ranch, the Century model predicted that forest clearance and conversion to pasture would cause an initial fall in soil C stocks, followed by a slow rise to levels exceeding those under native forest (Figure 9.3). The model predicted longer-term changes in soil C under pasture close to those inferred from the pasture chrono-sequence. Mean differences between simulated and observed values were about 9.32 g m-2 for total soil C (data not shown). After approximately 80 years under pasture cultivation, simulated results showed that soils of the Nova Vida Ranch chrono-sequence sequestered about 1.7 kg C m-2 in comparison to the levels presented by the soil under native forest (Figure 9.3). Bearing in mind that this figure is model derived, it is interesting to observe that soil C stocks increase, even for a period longer than 100 years. On the other hand, many studies related to soil C sequestration suggest that there is probably a limit for nutrient storage in a particular soil type under specific climatic and management conditions (Schlesinger, 2000; Schuman et al., 2002; Lal, 2003).

Figure 9.4 shows the simulated effects of climate change scenarios suggested by the Tyndall Center (Mitchell et al., 2003) on soil C stock dynamics in the Nova Vida forest-to-pasture chrono-sequence. Note that the same parameterization procedures were adopted in the three simulated conditions (current climate, HadCM3-A1F1, and DOE PCM-B1), except for the climatic variables temperature and precipitation, which were modified (Table 9.1) according to criteria discussed above.

Simulated results gave similar curve shapes for all three modeled situations, that is, an initial decline in soil C stock in the first 2 to 3 years following conversion from forest to pasture, and then a steady increase during pasture establishment. Small differences of simulated soil C dynamics between DOE PCM-B1 and HadCM3-A1F1 scenarios can be observed. For instance, in the first 16 to 17 years after deforestation,

Pasture age (years)

Figure 9.3 Simulated result of soil C content in the 0- to 20-cm layer at the forest to pasture chrono-sequence, Nova Vida Ranch, Rondonia State, Amazon region.

Pasture age (years)

Figure 9.3 Simulated result of soil C content in the 0- to 20-cm layer at the forest to pasture chrono-sequence, Nova Vida Ranch, Rondonia State, Amazon region.

the scenario DOE PCM-B1 presented slightly higher soil C stock results compared to the HadCM3-A1F1 scenario. Around year 20, the difference between those two modeled conditions disappeared. Moreover, after about 80 years of pasture cultivation, simulated results showed an inversion of the pattern presented in the early period, that is, soil C stock results were approximately 2% higher for the HadCM3-A1F1 scenario compared to the DOE PCM-B1 scenario (Figure 9.4).

According to Century model predictions, Nova Vida chrono-sequence soils under current climate conditions would store much more C in the 0 to 20 cm layer than the other two considered scenarios. Actually, simulated results applying weather data measured at the study area indicated that soil would sequester about 4160 g C m-2 after 80 years of continuous well-managed pasture cultivation, which is approximately 400 g C m-2 and 465 g C m-2 more than the HadCM3-A1F1 and DOE PCM-B1 scenarios, respectively (Figure 9.4).

4500

Actual weather data

2500

4000

Actual weather data

2500

4000

Pasture age (years)

Figure 9.4 Century-simulated scenarios of soil C stock dynamics at Nova Vida Ranch forest to pasturage chrono-sequence, applying Tyndall Center predictions.

2000

Pasture age (years)

Figure 9.4 Century-simulated scenarios of soil C stock dynamics at Nova Vida Ranch forest to pasturage chrono-sequence, applying Tyndall Center predictions.

Despite the enhancement in annual mean temperature of 1.3°C or 6.7°C (scenarios DOE PCM-B1 and HadCM3-A1F1, respectively), simulated results for those scenarios did not reflect in an increase of soil C stocks compared to levels in the actual weather data scenario. A plausible reason for this condition may be directly related to the Century model decomposition structure and concept. In the Century model, average monthly soil temperature near the soil surface is calculated using equations developed by Parton et al. (1987). These equations calculate maximum soil temperature as a function of the maximum air temperature and the canopy biomass (lower for high biomass), while minimum soil temperature is a function of the minimum air temperature and canopy biomass (higher for high biomass). The actual soil temperature used for decomposition and plant growth rate functions is the average of the minimum and maximum soil temperature (Metherell et al., 1993). Therefore, increasing temperature by 1.3°C or 6.7°C simulated decomposition rates, reducing the storage rates of C into the surface soil layer (Figure 9.4).

Moreover, it is interesting to observe that the former inference of decomposition levels would probably occur more intensively in the slow C pool, which is responsible for about 68% of the total C in the first 20 cm below the surface (Figure 9.5). As expected, independent of the soil C pool (active, slow, or passive) considered, simulated results followed the same pattern presented in Figure 9.4, that is, the highest soil C content showed for simulations applying actual weather data, and the lowest for simulations using climatic predictions from the DOE PCM-B1 scenario.

The decomposition of SOM in the Century model is assumed to be mediated with an associated loss of CO2 as a result of microbial respiration. The potential decomposition rate is reduced by multiplicative functions of soil moisture and soil temperature. Decomposition products flow into one

Pasture age (years)

Figure 9.5 Simulated soil C pools in the 0- to 20-cm layer, using actual weather data, HadCM3-A1F1 and DOE PCM-B1 scenarios, for forest-to-pasturage chrono-sequence conditions.

Pasture age (years)

Figure 9.5 Simulated soil C pools in the 0- to 20-cm layer, using actual weather data, HadCM3-A1F1 and DOE PCM-B1 scenarios, for forest-to-pasturage chrono-sequence conditions.

of the three SOM pools, each characterized by different maximum decomposition rates (Metherell et al., 1993). The active pool (Figure 9.5) represents soil microbes and microbial products, and has a turnover time of months to a few years, depending on the environment. Soil texture influences the turnover rate of the active SOM (higher rates for sandy soils) and the efficiency of stabilizing active into slow SOM. The slow pool in Figure 9.5 includes resistant plant material derived from the structural pool and soil-stabilized microbial products derived from the active and surface microbe pools. It has a turnover time of 20 to 50 years. Finally, the passive pool is very resistant to decomposition and includes physically and chemically stabilized SOM, and has a turnover time of 400 to 2000 years.

From the standpoint of soil C sequestration, the ideal situation is to store C in the passive pool, due to its stabilization state and long turnover time. Analyzing simulated results presented in Figure 9.5, it is possible to verify that about 29% of the total C content is in the passive pool. Moreover, simulated values are increasing steadily (the change appears linear because of the slow rate of change) throughout the simulation period, and are not maintaining a constant level as in the active pool.

The climate change scenarios impact other soil chemical, physical, and biological properties that we could not directly validate with measured data from the Nova Vida Ranch chrono-sequence. Another important aspect related to soil C dynamics in the Amazon region that we have not dealt with here is related to pasture management. Fearnside and Barbosa (1998) showed that trends in soil carbon were strongly influenced by pasture management. Sites that were judged to have been under bad management generally lost soil C, whereas sites under improved management had gained carbon. Trumbore et al. (1995) reported soil C losses in overgrazed pasture, but soil C gains from fertilized pasture in the Amazon region. Neill et al. (1997) suggested that degraded pastures with little grass cover probably will be less likely to accumulate soil C because inputs to soil organic C from pasture roots will be diminished, but that might not be true in

Table 9.2 Simulated Changes in Plant Productivity of Pasture at Nova Vida Ranch Using Current (Actual Data), HadCM3-A1F1 and DOE PCM-B1 Scenarios

Scenario Simulated Plant Productivity

Above Ground Below Ground

Current (actual data) DOE PCM-B1 HadCM3-A1F1

520 500 416

392 333

more vigorous secondary forest regrowth. Greater grazing intensity and soil damage from poor management would in all likelihood cause soil C and N losses. Similar processes that influence magnitude of annual SOM inputs also regulate the accumulation of C in soils of North American grasslands (Con-ant et al., 2001).

We have also simulated changes in above- and below-ground plant productivity for pastures at Nova Vida Ranch, using weather data from current (actual data) HadCM3-A1F1 and DOE PCM-B1 scenarios (Table 9.2). Modeled results showed a decrease in above- and below-ground productivity of 4% using DOE PCM-B1 data, and about 20% reduction using data from the HadCM3-A1F1 scenario compared to the plant productivity simulated for the current scenario (Table 9.2). Those simulated results suggest that climate change would cause a reduction in cattle stock rate (animals per hectare) for pastures in the Nova Vida Ranch.

Finally, as shown in the present study, modeling provides a flexible and powerful way to assess how different scenarios of climate and land use changes can affect soil C dynamics.

0 0

Post a comment