Abstract

The potential role of the impacts of land-cover changes (LCCs) in the Australian climate is investigated within the context of increasing CO2 concentrations and temperature. Specifically, it is explored if possible scenarios for LCC can moderate or amplify CO2-induced changes in climate over Australia. The January climate of Australia is simulated under three different land-cover-change scenarios using a high-resolution regional climate model. The land-cover-change scenarios include a steady-state land cover that is equivalent to current land cover, a low-reforestation scenario that recovers approximately 25% of the trees replaced by grasslands within the last 200 yr, and a high-reforestation scenario that recovers at least 75% of the deforested regions. The model was driven by boundary conditions taken from transitory climate simulations from a general circulation model that included two climate scenarios based on two projected scenarios of CO2 concentration increase. The results show that reforestation has the potential to reduce the projected increase in Australian temperatures in 2050 and 2100 by as much as 40% and 20%, respectively. This cooling effect, however, is highly localized and occurs only in regions of reforestation. The results therefore hint that the potential of reforestation to moderate the impact of global warming may be significantly limited by the spatial scale of reforestation. In terms of deforestation, results show that any future land clearing can exacerbate the projected warming in certain regions of Australia. Carbon-related variables are also analyzed and results show that changes in net CO2 flux may be influenced more by soil respiration than by photosynthesis. The results herein encourage studies on the inclusion of land-cover-change scenarios in future climate change projection simulations of the Australian climate.

1. Introduction

The significance of the impacts of historical land-cover change (LCC) on the present-day climate of Australia has been investigated by Narisma and Pitman (Narisma and Pitman 2003) and Pitman et al. (Pitman et al. 2004). Their results showed that LCC may account for a substantial part of the regional long-term weather changes over Australia in the last two centuries, including changes in temperature and rainfall. The significance of these results has established the important role of historical LCC in the evolution of the Australian climate to the present day. However, given projected changes in the Australian climate in the next 50 to 100 yr, LCC experiments must go beyond the impacts of historical changes in vegetation. Narisma and Pitman (Narisma and Pitman 2004) and Narisma et al. (Narisma et al. 2003) have shown that both changes in CO2 concentration and climate can affect Australian vegetation and hence affect the results of LCC experiments that omit these feedbacks. In this paper, we explore the impacts of LCC in the context of changing CO2 and climate and investigate the potential role of LCC in the future climate of Australia. Specifically, we explore if possible scenarios for future LCC moderate or amplify CO2-induced changes in climate over Australia.

There have been many studies on the impacts of future changes in CO2 and climate on vegetation (Betts et al. 1997; Cao and Woodward 1998; Woodward et al. 1998; Levis and Foley 1999; Betts et al. 2000; Levis et al. 2000; Bergengren et al. 2001; Claussen et al. 2001). Studies on the combined effects of LCC and climate change due to the enhanced greenhouse effects are less common (Costa and Foley 2000; Zhang et al. 2001; DeFries et al. 2002). The results of Costa and Foley (Costa and Foley 2000) showed that the effect of higher CO2 concentrations in the Amazon is to increase the average precipitation. However, the combined effects of deforestation and higher CO2 concentrations result in a reduction in the average rainfall. These results indicate the importance of taking into account the combined effects of LCC and higher CO2 concentrations with any corresponding global warming. This conclusion is supported by Zhang et al. (Zhang et al. 2001) who explored the compounding effects of deforestation and global warming and showed that LCC can potentially affect future climate projections.

To our knowledge, this is the first study on the impacts of future LCC on the Australian climate. In this paper, we explore the significance of the impacts of LCC on the projected increase in temperatures over Australia. Further, we investigate how carbon-related quantities are affected by LCC as temperatures and atmospheric CO2 concentrations increase over the next 100 yr. Understanding the role and significance of LCC in Australia’s changing climate is the first step to exploring the potential of LCC as a means to mitigate future climate changes that may be detrimental to the Australian environment.

2. Methodology

2.1. Model configuration

We used the Regional Atmospheric Modeling System (RAMS; Pielke et al. 1992; Liston and Pielke 2001) developed by the Colorado State University coupled to the General Energy and Mass Transport Model (GEMTM; Chen and Coughenor 1994; Eastman et al. 2001). RAMS is a flexible meteorological modeling system that has been extensively used to study the impact of LCC on weather and climate (see Pielke et al. 1998). More importantly, RAMS has been shown to simulate Australian atmospheric processes well (Peel et al. 2004). All RAMS simulations used the Kain and Fritsch (Kain and Fritsch 1993) convection scheme and the Chen and Cotton (Chen and Cotton 1987) shortwave and longwave radiation schemes. GEMTM is a dynamic plant model that simulates the interaction between the biosphere and atmosphere. At each time step, GEMTM calculates stomatal conductance as a function of relative humidity and CO2 concentration (Chen and Coughenor 1994). The photosynthetic rate is dependent on the atmospheric CO2 and vegetation temperature as well as on the photosynthetically active radiation and plant water potential. The coupled model, GEMRAMS, allows the vegetation to respond to changes in CO2 concentration and to any changes in climate, and it is for this reason that we chose to use this coupled model. We did not use the fifth-generation National Center for Atmospheric Research–Pennsylvania State University Mesoscale Model coupled with the Pleim-Xiu land surface scheme (MM5-PX) [used by Narisma and Pitman (2003) to investigate the impacts of historical LCC on Australian climate] because it lacks a dynamic vegetation scheme.

2.2. Data and experiments

The Australian Surveying and Land Information Group (AUSLIG; AUSLIG 1990) reconstructed past, including pre-European, land cover in Australia and the current land cover based on Landsat satellite imagery. The AUSLIG data consist of floristic types, which were mapped to the vegetation classifications in GEMRAMS (Table 1), except for Eucalypts, which were mapped into its own category (Peel et al. 2004). Based on these datasets, we extrapolated two land-cover scenarios, a low-recovery future (LCL) and a high-recovery future (LCH).

Table 1.

Mapping of the Australian floristic types to the vegetation categories in the models.

Mapping of the Australian floristic types to the vegetation categories in the models.
Mapping of the Australian floristic types to the vegetation categories in the models.

Using GEMRAMS, we performed 14 ensemble simulations of the Australian January climate over an 80 × 100 domain at 56-km grid spacing. Each ensemble simulation consists of four experiments with different land-cover scenarios using GEMRAMS. We chose to simulate the January climate because Narisma and Pitman (Narisma and Pitman 2003) have previously shown that there is a clear response to changes in land cover for this month. GEMRAMS was initialized and driven by boundary conditions taken from a transitory climate simulation of the Commonwealth Scientific and Industrial Research Organisation (CSIRO) Mark 2 atmosphere–ocean model (Watterson and Dix 2003). The CSIRO model has a spatial resolution of approximately 3.28° latitude and 5.68° longitude and includes nine vertical layers for the atmosphere. Climate simulations performed using the CSIRO model included A2 (high CO2 increase) and B2 (moderate CO2 increase) scenarios (see Nakicenovic et al. 2000). We updated boundary conditions for GEMRAMS every 12 h, which was found to be acceptable for simulations at a grid spacing of around 45 km by Denis et al. (Denis et al. 2003).

To provide baseline values, we performed experiments with past (pre-European) and current-day conditions in terms of climate, CO2 concentration, and land cover. For the projection experiments, we simulated future climate changes under the A2 and B2 scenarios (Nakicenovic et al. 2000) with three types of land-cover scenarios. The steady-state (SS) scenario retains the current (2000) land cover in both 2050 and 2100. LCL is a “low” reforestation scenario that recovers about 25% of the Eucalypt trees that were replaced by grasslands in the last 200 yr in the southeast and east (SE) and southwest (SW) regions of Australia. These are the areas most affected by historical LCC and its impacts on climate (Narisma and Pitman 2003). LCL also includes a region of land clearing, specifically shrubs replaced by grasslands, in the northeast (NE). The LCH land-cover scenario is a “high”-reforestation scenario that recovers at least 75% of the deforested regions in the SE and SW, but retains the land clearing in the NE. Figure 1 shows the difference between the SS land cover and the two land-cover scenarios. Areas in shades of orange denote regions of reforestation and those that are in shades of blue are areas of land clearing. Table 2 outlines the experiments performed in this paper.

Figure 1.

Difference in land cover between the steady-state and reforestation scenarios: (a) SS − LCL and (b) SS − LCH. Enclosed areas 1, 2, and 3 are the regions of LCC investigated in this study.

Figure 1.

Difference in land cover between the steady-state and reforestation scenarios: (a) SS − LCL and (b) SS − LCH. Enclosed areas 1, 2, and 3 are the regions of LCC investigated in this study.

Table 2.

Land-cover-change experiments that will explore the future impacts of LCC.

Land-cover-change experiments that will explore the future impacts of LCC.
Land-cover-change experiments that will explore the future impacts of LCC.

We focused our analyses on the simulated values for continental Australia and the three regions of LCC (SE, SW, and NE; boxed and marked in Figure 1b). These three regions will provide a guide to how reforestation and land clearing may affect future regional climate changes in Australia. In all experiments, the vegetation is allowed to respond to changes in CO2 concentration and climate, and leaf area index (LAI) is stabilized offline before the actual simulations to account for the lag between the physiological and structural response to CO2 and climate. A climate or CO2-driven change in LAI occurs at longer time scales (typically days to seasons) while the change in the stomatal conductance (gs) is almost immediate. We account for this lag by stabilizing the LAI before each simulation. GEMTM is run offline (uncoupled and without dynamic interaction with RAMS) using forcing data that were previously saved in a coupled January simulation but with seasonal and geographical variations in temperature, rainfall, and radiation imposed onto these data. We stabilize LAI iteratively by running the offline GEMTM for a year and computing for each grid point the difference between the 1-yr simulated and final LAI values. We start with an initial LAI value that is 0.25 smaller than the default value in the model (i.e., LAIinitial = LAIdefault – 0.25) to account for the possibility that the stable LAI may be less than the prescribed value. We iterate the offline GEMTM, increasing the LAI by 0.25 at the end of each annual cycle until the change in LAI between the simulated and initial LAI is less than or equal to 0.1 (see Figure 2).

Figure 2.

Schematic diagram of the LAI stabilization process used to account for the lag between the structural and physiological response to increases in CO2 and changes in climate. Figure taken from Narisma and Pitman (Narisma and Pitman 2004).

Figure 2.

Schematic diagram of the LAI stabilization process used to account for the lag between the structural and physiological response to increases in CO2 and changes in climate. Figure taken from Narisma and Pitman (Narisma and Pitman 2004).

3. Results and discussion

3.1. The impact of LCC on surface air temperature and rainfall

Figure 3 shows precipitation differences between SS and the two land-cover scenarios LCL and LCH for each emissions scenario and shows that there are no clear changes in rainfall patterns. Land-cover change, however, has noticeably affected surface air temperatures especially in areas where the changes in vegetation have occurred (Figure 4). Our results show that reforestation in LCH in the SE can potentially decrease the projected local warming in 2050 and 2100 by as much as 0.4°–0.5°C for all emissions scenario while the land clearing in NE has exacerbated the projected warming in NE by about 0.1°–0.2°C. The systematic temperature signal associated with each LCC experiment (Figure 4) is noteworthy and contrasts markedly with the apparently random changes in rainfall (Figure 3). However, the LCL has limited impacts on surface air temperature in SW because the change in land cover imposed is relatively small (Figure 1). In the LCH scenario, regional cooling of 0.4°–0.5°C occurs independent of the emission scenario or whether 2050 or 2100 is considered.

Figure 3.

Differences in simulated rainfall (mm day−1) between (a) SS and LCL in 2050A2, (b) SS and LCL in 2050B2, (c) SS and LCL in 2100A2, (d) SS and LCL in 2050B2, (e) SS and LCH in 2050A2, (f) SS and LCH in 2050B2, (g) SS and LCH in 2100A2, and (h) SS and LCH in 2050B2.

Figure 3.

Differences in simulated rainfall (mm day−1) between (a) SS and LCL in 2050A2, (b) SS and LCL in 2050B2, (c) SS and LCL in 2100A2, (d) SS and LCL in 2050B2, (e) SS and LCH in 2050A2, (f) SS and LCH in 2050B2, (g) SS and LCH in 2100A2, and (h) SS and LCH in 2050B2.

Figure 4.

Differences in simulated surface air temperature (°C) between (a) SS and LCL in 2050A2, (b) SS and LCL in 2050B2, (c) SS and LCL in 2100A2, (d) SS and LCL in 2050B2, (e) SS and LCH in 2050A2, (f) SS and LCH in 2050B2, (g) SS and LCH in 2100A2, and (h) SS and LCH in 2050B2.

Figure 4.

Differences in simulated surface air temperature (°C) between (a) SS and LCL in 2050A2, (b) SS and LCL in 2050B2, (c) SS and LCL in 2100A2, (d) SS and LCL in 2050B2, (e) SS and LCH in 2050A2, (f) SS and LCH in 2050B2, (g) SS and LCH in 2100A2, and (h) SS and LCH in 2050B2.

We calculated the changes in surface air temperature caused by LCC as a percentage of the increase in surface air temperatures between the current (2000) and projected temperatures with steady state (SS) land cover. To quantify, for example, the percentage cooling effect of reforestation in land-cover scenario LCL with respect to the warming for 2050A2, we have 100×ΔT(SS-LCL)2050A2T(Current-SS2050A2). This percentage would quantify the strength of the cooling effect of reforestation in SE and SW and the warming effect of land clearing in NE relative to the projected warming without reforestation (i.e., the SS land cover).

Figure 5 shows that both LCL and LCH decrease the warming projected, over areas of LCC, in SE and SW for the 2050A2 and 2050B2 scenarios by up to 20%–40%, but the impact in SW is very much larger in the LCH scenario, as would be expected (Figure 1). By 2100, the cooling effect of reforestation weakens relative to the projected increase in temperatures but still cools temperatures by 20% locally in LCL and over large areas of SE in LCH. Over NE, Figure 5 shows that the degree of warming caused by LCC, which exacerbates the warming due to the enhanced greenhouse effect, decreases from about 10% in 2050 to about 5% by 2100.

Figure 5.

Change in surface air temperature due to LCC expressed as a percentage of the change in temperature due to climate change for (a) 2050A2 LCL, (b) 2050B2 LCL, (c) 2100A2 LCL, (d) 2100B2 LCL, (e) 2050A2 LCH, (f) 2050B2 LCH, (g) 2100A2 LCH, and (h) 2100B2 LCH.

Figure 5.

Change in surface air temperature due to LCC expressed as a percentage of the change in temperature due to climate change for (a) 2050A2 LCL, (b) 2050B2 LCL, (c) 2100A2 LCL, (d) 2100B2 LCL, (e) 2050A2 LCH, (f) 2050B2 LCH, (g) 2100A2 LCH, and (h) 2100B2 LCH.

Figures 4 and 5 show that, for all scenarios, the impacts of LCC on surface air temperature are regional and are limited to the areas of LCC. The “localized” nature of the LCC impacts is evident in Figure 6, which shows area averages of temperature over continental Australia, SE, SW, and NE for each land-cover type and scenario. The area average over Australia shows no significant impacts on surface air temperature due to LCC (Figure 6a) (i.e., the three bars shown for A2 or B2 scenarios for SS, LCL, LCH are identical). Rather, the warming due to higher CO2 concentrations clearly dominates continental temperature (i.e., the size of all bars for 2050 are much larger than the bar for 2000). The three bars for A2 (or B2) scenarios showing the temperature for SS, LCL, and LCH are almost identical in both 2050 and 2100 with differences of less than 0.02°C between the three land-cover scenarios (Figure 6a). This contrasts with the change in temperature due to changes in CO2, which exceeds 1.9° (2000–2050) and 4.9°C (2000–2100). Clearly at a continental scale the signal due to increasing CO2 concentrations dominates any signal from LCC.

Figure 6.

Area averages for simulated surface air temperatures (°C) over (a) continental Australia, (b) southeast, (c) southwest, (d) and northeast regions of Australia.

Figure 6.

Area averages for simulated surface air temperatures (°C) over (a) continental Australia, (b) southeast, (c) southwest, (d) and northeast regions of Australia.

While the continental-scale impact of LCC is negligible, reforestation consistently decreases the CO2-induced warming over areas of LCC in the SE and SW regions in 2050 and 2100 for all scenarios. The area-averaged cooling is much smaller than the CO2-induced increase in temperatures between the current and projected climates. However, the cooling is not negligible and the magnitude of the cooling effect of LCH in all scenarios in SE is more than the simulated warming (about 0.11°C) in this region in the last 200 years (Figure 6b). In the B2 scenarios, for example, reforestation cools temperatures by about 0.15°C in 2050 and 2100. For the SW region, the cooling effect of reforestation is almost half of the simulated warming over the last two centuries (Figure 6c). The warming in the NE due to land clearing is small compared to the increase in temperatures due to the enhanced greenhouse effect.

3.2. Mechanisms behind the impact of LCC on Australian temperatures

The cooling effect of reforestation in SE and SW is mainly due to an increase in latent heat flux (λE) resulting from an increase in LAI (Figures 7 and 8). In the reforested areas in SE and SW, the LAI values are higher by as much as 2.0–2.5 m2 m−2 in LCL and LCH than in the SS land cover (Figure 7). The higher LAI led to an increase in the latent heat flux in these reforested regions in both LCL and LCH by about 100–150 W m−2 compared with the SS latent heat flux in all emissions scenarios (Figure 8). A comparison of the change in LAI (Figure 7) and latent heat flux (Figure 8) shows a very strong and consistent relationship across all scenarios. Figure 8 also shows some changes in latent heat flux in regions remote from LCC, but these are small and are likely caused by the random changes in simulated precipitation (Figure 3).

Figure 7.

Differences in simulated LAI (m2 m−2) between (a) SS and LCL in 2050A2, (b) SS and LCL in 2050B2, (c) SS and LCL in 2100A2, (d) SS and LCL in 2050B2, (e) SS and LCH in 2050A2, (f) SS and LCH in 2050B2, (g) SS and LCH in 2100A2, and (h) SS and LCH in 2050B2.

Figure 7.

Differences in simulated LAI (m2 m−2) between (a) SS and LCL in 2050A2, (b) SS and LCL in 2050B2, (c) SS and LCL in 2100A2, (d) SS and LCL in 2050B2, (e) SS and LCH in 2050A2, (f) SS and LCH in 2050B2, (g) SS and LCH in 2100A2, and (h) SS and LCH in 2050B2.

Figure 8.

Differences in simulated latent heat flux (W m−2) between (a) SS and LCL in 2050A2, (b) SS and LCL in 2050B2, (c) SS and LCL in 2100A2, (d) SS and LCL in 2050B2, (e) SS and LCH in 2050A2, (f) SS and LCH in 2050B2, (g) SS and LCH in 2100A2, and (h) SS and LCH in 2050B2.

Figure 8.

Differences in simulated latent heat flux (W m−2) between (a) SS and LCL in 2050A2, (b) SS and LCL in 2050B2, (c) SS and LCL in 2100A2, (d) SS and LCL in 2050B2, (e) SS and LCH in 2050A2, (f) SS and LCH in 2050B2, (g) SS and LCH in 2100A2, and (h) SS and LCH in 2050B2.

In the NE region, Figure 7 shows that the LAI for 2050A2, 2050B2, and 2100B2 in the LCL and LCH land-cover scenarios are lower (by up to 2 m2 m−2) compared to the LAI in the SS land-cover scenario. The lower LAI due to LCC has resulted in a lower latent heat flux of about 2.5–5 W m−2, which caused the slight increase in temperature (Figure 4) and exacerbated the projected warming due to increasing CO2 over this region. By 2100A2, Figure 7g shows that there is a large reduction in LAI in the LCL and LCH scenarios, which may be due to an increase in water stress brought about by the much warmer temperatures. Figure 8g shows that despite the large decrease in LAI, the latent heat flux in the LCL and LCH land-cover scenarios is higher than the latent heat flux in the SS land-cover scenario, an effect driven by higher evaporative demand. The simulated warming in temperatures due to LCC in the NE can be explained by the reduction in albedo (Figure 9), which is a consequence of the large decrease in LAI, which reduces canopy interception of incoming solar radiation.

Figure 9.

Differences in simulated albedo between (a) SS and LCL in 2050A2, (b) SS and LCL in 2050 B2, (c) SS and LCL in 2100 A2, (d) SS and LCL in 2050B2, (e) SS and LCH in 2050A2, (f) SS and LCH in 2050 B2, (g) SS and LCH in 2100 A2, and (h) SS and LCH in 2050B2.

Figure 9.

Differences in simulated albedo between (a) SS and LCL in 2050A2, (b) SS and LCL in 2050 B2, (c) SS and LCL in 2100 A2, (d) SS and LCL in 2050B2, (e) SS and LCH in 2050A2, (f) SS and LCH in 2050 B2, (g) SS and LCH in 2100 A2, and (h) SS and LCH in 2050B2.

The area averages in LAI (Figure 10) show an increase in LAI for the LCL and LCH in all scenarios even at continental scale. The increase in LAI area averaged over Australia (Figure 10a), however, is small (maximum of about 0.1 m2 m−2) compared to the regional-scale changes and has not considerably affected continental averages for latent heat flux (Figure 11). In contrast, the LAI in SE (Figure 10b) has increased by up to 0.6–0.65 m2 m−2 in LCH for both A2 and B2 scenarios in 2050 and 2100. The higher LAI in SE for the LCL and LCH scenarios increases the latent heat flux by about 9 and 18 W m−2, respectively (Figure 11b). Over SW, the LAI for the reforested regions increases to a maximum of 0.98 m2 m−2 compared to the SS LAI (Figure 10c) resulting in a maximum increase of almost 40 W m−2 in latent heat flux (Figure 11c). The reduction in LAI caused by land clearing over NE is consistent through all scenarios but is most prominent in 2100A2 where LAI has decreased by almost 2.5 m2 m−2 (Figure 10d). This has changed the sign of the change in albedo in this region (Figure 12g) and has reduced the albedo by about 2% increasing temperatures (Figure 6d).

Figure 10.

Area averages for LAI (m2 m−2) over (a) continental Australia, (b) southeast, (c) southwest, and (d) northeast regions of Australia.

Figure 10.

Area averages for LAI (m2 m−2) over (a) continental Australia, (b) southeast, (c) southwest, and (d) northeast regions of Australia.

Figure 11.

Area averages for latent heat flux (W m−2) over (a) continental Australia, (b) southeast, (c) southwest, and (d) northeast regions of Australia.

Figure 11.

Area averages for latent heat flux (W m−2) over (a) continental Australia, (b) southeast, (c) southwest, and (d) northeast regions of Australia.

Figure 12.

Area averages for albedo over (a) continental Australia, (b) southeast, (c) southwest, and (d) northeast regions of Australia.

Figure 12.

Area averages for albedo over (a) continental Australia, (b) southeast, (c) southwest, and (d) northeast regions of Australia.

3.3. Isolating the role of LAI in the simulated changes in temperature

The impact of LCC on surface air temperature is largely governed by the physical changes in LAI. For the reforested regions SE and SW, the increase in LAI for all scenarios has increased latent heat flux and has led to a reduction in the projected warming in the SS land-cover scenario. The land clearing imposed on NE has reduced LAI, which intensified the warming in the SS scenario via two different mechanisms. The temperature increase due to LCC in 2050 is due to the reduction in latent heat flux caused by the decrease in LAI. In the 2100A2 scenario, the warming was caused by an increase in net radiation brought about by the decrease in albedo due to the sharp decrease in LAI.

We isolated the impacts of LAI to explore the dominance of this structural feedback over the physiological impact, that is, the reduction in stomatal conductance associated with higher CO2 concentrations (Betts et al. 1997; Betts et al. 2000; DeFries et al. 2002). Using the LCH land cover for the A2 scenarios of 2050 and 2100, we allowed LAI to respond to the increase in CO2 but kept the CO2 sensed by the stomates to current values of 369 ppmv. Results show that the simulated changes in temperature due to LAI in the isolation experiments (Figures 13a and 13c) are very similar to the simulated changes in Figures 4e and 4g. We calculated the temperature changes in Figures 4e and 4g that were accounted for by the changes in LAI alone (i.e., %=100×(ΔtemperatureΔLAI/ΔtemperatureΔLAI+Δgs). Figures 13b and 13d show that changes in LAI account for most of the impacts simulated in the LCC regions.

Figure 13.

Change in surface air temperature (°C) due to change in LAI only, i.e., without physiological feedback for (a) 2050A2 LCH and (c) 2100A2 LCH and percentage of the temperature changes in Figures 4e and 4g that are accounted for by the changes in LAI alone for (b) 2050A2 LCH and (d) 2100A2 LCH.

Figure 13.

Change in surface air temperature (°C) due to change in LAI only, i.e., without physiological feedback for (a) 2050A2 LCH and (c) 2100A2 LCH and percentage of the temperature changes in Figures 4e and 4g that are accounted for by the changes in LAI alone for (b) 2050A2 LCH and (d) 2100A2 LCH.

3.4. The impact of LCC on carbon-related variables

Given the importance under greenhouse mitigation strategies of changes in net carbon stores, it is useful to explore how carbon-related quantities may change in the future. In our experiments, photosynthesis has generally increased for LCL and LCH in all regions (SE, SW, NE, and AUS) in all scenarios irrespective of the type of LCC (Figure 14). As with the other variables, the changes at the continental scale are small compared to the changes at the regional scale due to CO2 fertilization. The larger increases in photosynthesis occur over the reforested regions of SE and SW (Figure 14). These increases in photosynthesis are due to the increase in simulated stomatal conductance (Figure 15) due to CO2 fertilization. In the NE region, grasslands in the LCC scenarios have also responded positively to the increase in CO2 concentrations. As a result the simulated photosynthesis with LCC (grasslands) is consistently higher than the SS values for shrubs for all scenarios. The more positive response of grass compared with shrubs to CO2 concentration has been previously shown (Narisma and Pitman 2004).

Figure 14.

Differences in simulated photosynthesis (μmol CO2 m−2 s−1) between (a) SS and LCL in 2050A2, (b) SS and LCL in 2050B2, (c) SS and LCL in 2100A2, (d) SS and LCL in 2050B2, (e) SS and LCH in 2050A2, (f) SS and LCH in 2050B2, (g) SS and LCH in 2100A2, and (h) SS and LCH in 2050B2.

Figure 14.

Differences in simulated photosynthesis (μmol CO2 m−2 s−1) between (a) SS and LCL in 2050A2, (b) SS and LCL in 2050B2, (c) SS and LCL in 2100A2, (d) SS and LCL in 2050B2, (e) SS and LCH in 2050A2, (f) SS and LCH in 2050B2, (g) SS and LCH in 2100A2, and (h) SS and LCH in 2050B2.

Figure 15.

Differences in simulated stomatal conductance (m s−1) between (a) SS and LCL in 2050A2, (b) SS and LCL in 2050B2, (c) SS and LCL in 2100A2, (d) SS and LCL in 2050B2, (e) SS and LCH in 2050A2, (f) SS and LCH in 2050B2, (g) SS and LCH in 2100A2, and (h) SS and LCH in 2050B2.

Figure 15.

Differences in simulated stomatal conductance (m s−1) between (a) SS and LCL in 2050A2, (b) SS and LCL in 2050B2, (c) SS and LCL in 2100A2, (d) SS and LCL in 2050B2, (e) SS and LCH in 2050A2, (f) SS and LCH in 2050B2, (g) SS and LCH in 2100A2, and (h) SS and LCH in 2050B2.

In terms of net CO2 fluxes (μmol m−2 s−1), continental Australia (Figure 16a) and NE (Figure 16d) are net sources of carbon in our simulations while SE and SW are carbon sinks (Figures 16b and 16c, respectively). With reforestation, however, the net carbon uptake in the SE and SW decreases. Area averages for soil respiration show that the reduction in net CO2 flux in SE and SW may be explained by an increase in soil respiration (Figures 17b and 17c). For each scenario, the change in land cover from grass to trees over SE and SW has increased soil respiration and, as a consequence, decreased the net carbon uptake in these regions. The increase in soil respiration is driven by the combined effects of temperature and soil moisture changes (Figure 18) [see Chen and Coughenor 1994, Equation (57)]. Despite the cooling effect of reforestation, the increase in soil moisture (Figure 18) combined with warmer temperatures (Figure 6) due to the enhanced greenhouse effect has increased soil respiration in the reforested regions. This indicates that soil moisture may be a major driving factor in carbon calculations for future carbon emissions from the Australian biosphere. The strong influence of soil moisture is clear in 2100A2 over the SE and SW regions where despite warming (Figures 4b and 4c), the sharp decrease in soil moisture (Figures 18b and 18c) has led to reductions in soil respiration (Figures 17b and 17c). We emphasize that these are only January simulations and our results are likely to be highly dependent on how these processes are parameterized and it is not reasonable to extrapolate these findings to an annual estimate of the net carbon balance over Australia or elsewhere.

Figure 16.

Area averages for net CO2 flux (μmol CO2 m−2 s−1) over (a) continental Australia, (b) southeast, (c) southwest, and (d) northeast regions of Australia.

Figure 16.

Area averages for net CO2 flux (μmol CO2 m−2 s−1) over (a) continental Australia, (b) southeast, (c) southwest, and (d) northeast regions of Australia.

Figure 17.

Area averages for soil respiration (μmol CO2 m−2 s−1) over (a) continental Australia, (b) southeast, (c) southwest, and (d) northeast regions of Australia.

Figure 17.

Area averages for soil respiration (μmol CO2 m−2 s−1) over (a) continental Australia, (b) southeast, (c) southwest, and (d) northeast regions of Australia.

Figure 18.

Area averages for soil moisture content (topmost layer) over (a) continental Australia, (b) southeast, (c) southwest, and (d) northeast regions of Australia.

Figure 18.

Area averages for soil moisture content (topmost layer) over (a) continental Australia, (b) southeast, (c) southwest, and (d) northeast regions of Australia.

Land clearing in NE has increased the potential of this region to be a net carbon source (Figure 16d). The higher negative CO2 flux in NE cannot be fully explained by soil respiration or photosynthesis. Figure 17d shows that there is a reduction in soil respiration in all scenarios except for 2100A2 and it has been shown earlier that there is a consistent increase in photosynthesis in all scenarios (Figure 14). One explanation for the higher carbon emissions in the NE is an increase in plant respiration, and this variable has not been explored in this paper but will be investigated further in the future. The increase in soil respiration for 2100A2 in the NE is largely due to the warming in temperature (Figure 6d) rather than soil moisture change.

4. Summary and conclusions

This paper has explored the role of LCC on the future climate of Australia. Our results showed that reforestation in the SE and SW has the potential to reduce the warming caused by the enhanced greenhouse effect by as much as 40% under the A2 and B2 scenarios by 2050, but this cooling effect declines to 20% by 2100. The reduction in temperatures is the result of the increase in LAI, which led to a corresponding increase in latent heat flux. This cooling effect of reforestation is localized and there were no simulated changes in temperature over regions without LCC. The area-averaged temperature for continental Australia also showed no significant changes in temperature due to LCC. Hence, any cooling effect of LCC via reforestation would depend on the spatial scale of reforestation and, at least to first order, the cooling effect appears linearly related to the spatial scale of reforestation. This, unfortunately, significantly limits the potential to ameliorate CO2-induced warming in the long-term since the scale necessary for reforestation is unlikely to be realistic (indeed, the LCL scenario is likely overly optimistic).

Our results also showed that land clearing may exacerbate the projected warming in temperatures in NE. This warming, however, is relatively small compared to the increase in temperatures projected for the region under either A2 or B2 scenarios. The warming was explained by two different mechanisms arising from the reduction in LAI. In 2050A2, 2050B2, and 2100B2, the decrease in LAI due to LCC has reduced latent flux and, as a consequence, increased temperatures. By 2100A2, the large reduction in LAI, probably due to plant water stress as reflected in the decrease in soil moisture for this region, has reduced albedo and hence increased the net surface radiation, leading to a warming in temperatures. This result may be particularly model dependent since it relies on the simulation of soil moisture, the relationship between soil moisture and LAI, and the relationship between LAI and canopy albedo.

We also analyzed carbon-related variables, and our results showed that despite the increase in photosynthesis, the net CO2 flux was mainly affected by the increase in soil respiration. Soil respiration was determined by the combined effects of soil moisture and temperature, and our results indicate that soil moisture can be a major driving factor in carbon calculations for future biosphere carbon emissions. In areas where there were significant reductions in soil moisture, there were corresponding decreases in the simulated soil respiration despite increases in temperature. Our results therefore imply that soil moisture deserves equal importance with temperature in soil respiration calculations when accounting for carbon fluxes in the Australian biosphere. GEMTM used a parameterization of soil respiration used by Chen and Coughenor [Chen and Coughenor 1994, Equation (57)]. This parameterization is noted by Chen and Coughenor (Chen and Coughenor 1994) to warrant further experimental and modeling work and represents a first approximation of soil respiration. We therefore recognize uncertainties associated with the results reported in this paper.

There are other limitations in our experiments. For one, we performed 1-month simulations for the January climate, which limits a full dynamic feedback between the biosphere, atmospheric processes, and carbon. Our experiments also do not involve feedbacks to a larger global circulation. Our analyses are also limited by using two projected land-cover scenarios only. Other serious limitations come from model dependencies and uncertainties relating to the parameterization of land surface processes. The simulation of carbon processes is at the cutting edge of current capacity and the outcomes are highly uncertain (cf., e.g., Cox et al. 2000 and Friedlingstein et al. 2001). The impact of LCC and increasing CO2 concentrations on photosynthesis and soil respiration is clearly a function of a vast array of poorly represented quantities (e.g., soil moisture) and nonrepresented quantities (e.g., soil nutrients and changes in atmospheric deposition of nitrogen and sulfur). However, our demonstration that changes in LAI largely explained the role of LCC in moderating temperature rises suggests that further effort to understanding how climate might affect LAI would be one way of reducing current uncertainties.

Overall, our results are therefore no more than illustrative and should be seen as an initial attempt to explore the role that reforestation might have in moderating the impact of increasing CO2 concentrations over Australia. With this in mind, we find that reforestation does have the potential to offset global warming but only locally where large-scale reforestation occurs. The reductions in warming caused by increasing CO2 are quite significant (20%–40%) to 2050, but decline in importance as warming intensifies to 2100. The lack of remote effects caused by reforestation in our experiments substantially moderates the potential of reforestation as a means to offset greenhouse-induced warming over Australia. Specifically, the scale of reforestation in LCL is likely overly optimistic and given the time scales required for the forests to grow, coupled with the major uncertainties in our understanding of the biophysical system, large-scale reforestation does not appear to be a significant method of reducing warming due to greenhouse. It may be that seasonal simulations, or simulations over many years, will show remote effects or an impact of reforestation on rainfall remote from the regions of change, and these could change this conclusion. We therefore see the extension of our work to multiyear simulations as a priority.

Irrespective of the lack of remote impacts of LCC on temperature over Australia, the local impacts were substantial under both A2 and B2 scenarios to 2050. Therefore, while we do not argue for a large-scale reforestation project over Australia (the uncertainties are simply too large) we do conclude that LCC scenarios should be included in future projections of the Australian climate since these would add realism and regional detail to these projections. We note that a very substantial model and scenario development exercise is required to build confidence in the reliability of climate projections that include all significant forcing terms on the future climate of Australia.

Acknowledgments

We thank Ian Watterson at CSIRO for providing the boundary conditions from the CSIRO GCM. G. T. Narisma was supported by a MUIPGRA scholarship.

REFERENCES

REFERENCES
AUSLIG
1990
.
Atlas of Australian Resources: Vegetation. Australian Surveying and Land Information Group, Commonwealth of Australia, 64 pp
.
Bergengren
,
J. C.
,
S. L.
Thompson
,
D.
Pollard
, and
R. M.
Deconto
.
2001
.
Modeling global climate–vegetation interactions in a doubled CO2 world.
Climatic Change
50
:
31
75
.
Betts
,
R. A.
,
P. M.
Cox
,
S. E.
Lee
, and
F. I.
Woodward
.
1997
.
Contrasting physiological and structural vegetation feedbacks in climate change simulations.
Nature
387
:
796
799
.
Betts
,
R. A.
,
P. M.
Cox
, and
F. I.
Woodward
.
2000
.
Simulated response of potential vegetation to doubled-CO2 climate change and feedbacks on near-surface temperature.
Global Ecol. Biogeogr.
9
:
171
180
.
Cao
,
M.
and
F. I.
Woodward
.
1998
.
Dynamic responses of terrestrial ecosystem carbon cycling to global climate change.
Nature
393
:
249
252
.
Chen
,
C.
and
W. R.
Cotton
.
1987
.
The physics of the marine strato-cumulus-capped mixed layer.
J. Atmos. Sci.
44
:
2951
2977
.
Chen
,
D-X.
and
M. B.
Coughenor
.
1994
.
GEMTM: A general model for energy and mass transfer of land surfaces and its application at the FIFE sites.
Agric. For. Meteor.
68
:
145
171
.
Claussen
,
M.
,
V.
Brovkin
, and
A.
Ganopolski
.
2001
.
Biogeophysical versus biogeochemical feedbacks of large-scale land cover change.
Geophys. Res. Lett.
28
:
1011
1014
.
Costa
,
M. H.
and
J. A.
Foley
.
2000
.
Combined effects of deforestation and doubled atmospheric CO2 concentrations on the climate of Amazonia.
J. Climate
13
:
18
34
.
Cox
,
P. M.
,
R. A.
Betts
,
C. D.
Jones
,
S. A.
Spall
, and
I. J.
Totterdell
.
2000
.
Acceleration of global warming due to carbon-cycle feedbacks in a coupled climate model.
Nature
408
:
184
187
.
DeFries
,
R. S.
,
L.
Bounoua
, and
G.
Collatz
.
2002
.
Human modification of the landscape and surface climate in the next fifty years.
Global Change Biol.
8
:
438
458
.
Denis
,
B.
,
R.
Laprise
, and
D.
Caya
.
2003
.
Sensitivity of a regional climate model to the resolution of the lateral boundary conditions.
Climate Dyn.
20
.
doi:10.1007/s00382-002-0264-6
.
Eastman
,
J. L.
,
M. B.
Coughenor
, and
R. A.
Pielke
Sr.
.
2001
.
The regional effects of CO2 and landscape change using a coupled plant and meteorological model.
Global Change Biol.
7
:
797
815
.
Friedlingstein
,
P.
,
L.
Bopp
,
P.
Ciais
,
J-L.
Dufresne
,
L.
Fairhead
,
H.
LeTreut
,
P.
Monfray
, and
J.
Orr
.
2001
.
Positive feedback between future climate change and the carbon cycle.
Geophys. Res. Lett.
28
:
1543
1546
.
Kain
,
J. S.
and
J. M.
Fritsch
.
1993
.
Convective parameterization for mesoscale models: The Kain–Fritsch scheme. The Representation of Cumulus Convection in Numerical Models, Meteor. Monogr., No. 24, Amer. Meteor. Soc., 165–170
.
Levis
,
S.
and
J. A.
Foley
.
1999
.
Potential high-latitude vegetation feedbacks on CO2 induced climate change.
Geophys. Res. Lett.
26
:
747
750
.
Levis
,
S.
,
J. A.
Foley
, and
D.
Pollard
.
2000
.
Large-scale vegetation feedbacks on a doubled CO2 climate.
J. Climate
13
:
1313
1325
.
Liston
,
G. E.
and
R. A.
Pielke
.
2001
.
A climate version of the Regional Atmospheric Modeling System.
Theor. Appl. Climatol.
68
:
155
173
.
Nakicenovic
,
N.
Coauthors
2000
.
IPCC Special Report on Emissions Scenarios.
Cambridge University Press, 599 pp
.
Narisma
,
G. T.
and
A. J.
Pitman
.
2003
.
The impact of 200 years of land cover change on the Australian near-surface climate.
J. Hydrometeor.
4
:
424
436
.
Narisma
,
G. T.
and
A. J.
Pitman
.
2004
.
The effect of including biospheric feedbacks on the impact of land cover change over Australia.
Earth Interactions
8
.
[Available online at http://EarthInteractions.org.]
.
Narisma
,
G. T.
,
A. J.
Pitman
,
J.
Eastman
,
I. G.
Watterson
,
R.
Pielke
Sr.
, and
A.
Beltrán-Przekurat
.
2003
.
The role of biospheric feedbacks in the simulation of the impact of historical land cover change on the Australian January climate.
Geophys. Res. Lett.
30
.
2168, doi:10.1029/2003GL018261
.
Peel
,
D.
,
A. J.
Pitman
,
L.
Hughes
, and
G. T.
Narisma
.
2004
.
The impact of an explicit representation of Eucalyptus on the simulation of the January climate of Australia.
Environ. Modell. Software
20
:
595
612
.
Pielke
,
R. A.
Coauthors
1992
.
A comprehensive meteorological modeling system—RAMS.
Meteor. Atmos. Phys.
49
:
69
91
.
Pielke
Sr.,
R. A.
,
R.
Avissar
,
M.
Raupach
,
A. J.
Dolman
,
X.
Zeng
, and
A. S.
Denning
.
1998
.
Interactions between the atmosphere and terrestrial ecosystems: Influence on weather and climate.
Global Change Biol.
4
:
461
475
.
Pitman
,
A. J.
,
G. T.
Narisma
,
R. A.
Pielke
Sr.
, and
N. J.
Holbrook
.
2004
.
The impact of land cover change on the climate of southwest Western Australia.
J. Geophys. Res.
109
.
doi:10.11029/2003JD004347
.
Watterson
,
I.
and
M.
Dix
.
2003
.
Simulated changes due to global warming in daily precipitation means and extremes and their interpretation using the gamma distribution.
J. Geophys. Res.
108
.
4397, doi:10.1029/2002JD002928
.
Woodward
,
F. I.
,
M. R.
Lomas
, and
R. A.
Betts
.
1998
.
Vegetation–climate feedbacks in a greenhouse world.
Philos. Trans. Roy. Soc. London
B353
:
29
39
.
Zhang
,
H.
,
A.
Henderson-Sellers
, and
K.
McGuffie
.
2001
.
The compounding effects of tropical deforestation and greenhouse warming on climate.
Climatic Change
49
:
309
338
.

Footnotes

* Corresponding author address: Gemma T. Narisma, Center for Sustainability and the Global Environment, Nelson Institute for Environmental Studies, University of Wisconsin—Madison, 1710 University Ave., Madison, WI 53726. narisma@wisc.edu