Tropical Deforestation
Over the last several years, our group has explored a wide variety of climatic effects of tropical deforestation as well as tropical land use and land cover change. The tropics are an important part of the global climate system, and large-scale changes are contributing to climate change through various feedbacks between the atmosphere and the biosphere. Changes to the climate are occurring locally, regionally, and perhaps even globally.
Our group studies the influence of tropical land use and land cover change on the cliamte from local to global scales. We focus on energy, water, and caron dioxide exchanges between the biosphere and the atmosphere. In particular, we have investigated how changes in boundary layer energy can translate to larger-scale changes in convection, and how these changes may influence the extratropics by way of atmospheric teleconnections. While many studies have identified a cause-and-effect relationship between tropical deforestation and extratropical climate change, we have explicitly identified the processes that change, and more importantly, why they change. By examining changes to the climate dynamics, a clear picture of the teleconnection process emerges.
The following flowchart describes the surface and regional climate response to tropical deforestation. Overlall, there is a large surface warming due to the reduction in latent cooling associated with removal of the vegetation. Although the surface albedo increases, the decrease in surface net radiation is more than offset by the large reduction in the latent heat flux. The change in the surface and boundary layer energy partitioning and hydrologic cycle leads to regional scale changes that have an impact from the surface up through the troposphere. Model results suggest a 28% decrease in latent heat flux, a 1.0 K increase in surface temperature and a 20% decrease in precipitation (convective in nature).
The seasonal temperature response to large-scale deforestation is shown for all tropical forest regions below. As discussed, reductions in latent cooling contribute to an overall surface warming for the regions where deforestation occurs, but what about the large warming over Siberia that is most evident in Northern Hemisphere winter and spring?
The precipitation changes associated with the deforestation are most pronounced in December-January-February. The following figure shows the precipitation change for DJF. Note the strong reduction in precipitation (primarily the deep convective type) over South America as well as the repartitioning of precipitation to the south of the basin.
To better understand the regional climate response to tropical deforestation, one can explore the individual components of the tropospheric energy budget from the surface up through the atmosphere. The following three-dimensional energy budget equation includes terms (from left to right) for latent heat content due to phase changes, convergence of radiative heat, convergence of internal, gravitational potential, and latent energy fluxes).
Use of this equation with model output from tropical deforestation experiments suggests that there is a 20% reduction in convective precipitation. The equation also illuminates where this energy is going - deep convection is reduced in favor of shallow convection over the regions of deforestation, and deep convection is enhanced in areas down stream of the deforested regions as energy is imported. The following cartoon illustrates how energy is moved around between deforested and forested regions as well as what kind of energy changes are occuring at different heights in the atmosphere.
The repartitioning of shallow and deep convection as discussed above has a strong effect on tropical outflow. That is, the shallow convection is enhanced at the expense of deep convection, and the strength and positioning of the centers of deep convection are altered such that the interaction between the tropics and the extratropics are affected.
The following figure shows how tropical deforestation weakens the tropical outflow (especially over South America) by up to 20%. This reduction disrupts the tropical-to-extratropical flow of energy. The top panel shows the control case when tropical outflow dominates the three tropical heating centers. The bottom panel illustrates how deforestation weakens the tropical outflow and affects the energy transfer out of the tropics towards the poles.
In DJF, the poleward transport is strongest to the Northern Hemisphere. The change in the outflow is manifested in zonal velocity changes to both the sub-tropical and high-latitude jets. The following figure illustrates how these jets change for the month of January, when the teleconnection is strongest. The top panel shows the control case and the bottom panel shows the zonal wind changes with deforestation.
While the following figure shows a cause-and-effect relationship between tropical deforestation and extratropical dynamical changes, the real question is how does the signal propagate out of the tropics to higher latitudes? It turns out that the weakening of the tropical outflow causes modulation of the tropical-to-extratropical large-scale flow via an anomalous Rossby Wave train that eminates from the tropics and travels northeastward. The following figure shows the change in the rossby wave train with deforestation (with the 250 hPa stream function).
Pockets of changes in the sign of the stream function northeastward from South America provide the dynamical link between the tropics and extratropics. But this still doesn't explain the large warming throughout part of Siberia.
It turns out that changes to the Rossby (or planetary) wave dynamics results in changes in the storm tracks in the northern high-latitudes. The following figure illustrates how deforestation leads to Northern Hemisphere changes in the storm track intensity and positioning. The storm track is determined from band-pass filtered daily geopotential heights to isolate 2-8 day synoptic-scale events representing the storm tracks. The top panel shows the control case and the bottom panel shows the difference due to deforestation. Note the significant changes over Scandinavia and part of Siberia.
But the question still remains - how do changes in the storm track affect the warming? Before we answer that, we need to look at how synoptic-scale events (eddies) contribute to the mean flow. It turns out that synoptic eddies provide a poleward transport of energy (temperature and momentum). The following figure illustrates the change in the zonally-averaged eddy momentum flux across Eurasia. Note the northward shift and intensification of the eddy momentum flux. This change is a due to the previously discussed modification to the storm tracks. These data have been band-pass filtered to isolate the synoptic-scale eddies.
However, this still does not explain the observed warming throughout Siberia. Recall the traditional three-cell model of the atmosphere's meridional circulation. There is a Hadley cell operating between the tropics and the sub-tropics, a Polar cell operating at very high latitude, and in between, a Ferrel cell. The Ferrel cell is unique because it is a thermally indirect cell that drives cold air aloft and transports it towards the equator. That is, it is indirect because it transports energy from cold to warm regions. Interestingly, the Ferrel is a byproduct of a strong poleward transport of energy by synotpic eddies. The following cartoon illustrates this relationship.
So the idea is that if the synoptic eddies are affected by the storm track and anomalous Rossby wave activity, and furthermore, that the synotpic eddies affect the Ferrel cell dynamics, then it stands to reason that changes in the position and intensity of the synoptic eddie momentum flux will lead to changes in the Ferrel cell position and intensity. This is exactly what the above cartoon illustrates. That is, the mean meridional circulation is modified by eddy heat and momentum fluxes and the changes result in a northward shift and intensification of the Ferrel cell.
A key feature of the descending branch of the Ferrel cell is that as the air descends, it adiabatically warms. It is this adibatic warming with descent that leads to an overall lower-tropospheric warming around the latitude at which the large Siberian warming is observed. The following figure shows this warming as the zonally-averaged temperature over Eurasia in January. Note the very large surface warming that extends upwards of 700 mb. This large warming signature can have all sorts of climatic ramifications.
One such ramification of this large tropospheric warming over Siberia is related to land-atmosphere feedbacks - interestingly, some of the very same types of atmosphere-biosphere feedbacks driving the initial climate change in the tropics. The next figure shows the low-level cloud cover and surface albedo changes due to tropical deforestation.
As expected, the low-level tropospheric warming causes an expansion of the planetary boundary layer and a decrease in low-level cloud cover fraction over part of the region where the warming is greatest. The lower panel shows a decrease in the surface albedo consistant with a higher near-surface temperatures contributing to snow melt, thus exposing the bare soil beneath. Both reductions in low-level cloud cover and surface albedo both contribute to increasing the amount of net radiation at the surface. These two positive feedbacks contribute to further warming. That is, the initial warming due to an increase in descending air that is adiabatically warming, is amplified by these regional-scale land-atmosphere feedbacks. The resulting warming in Asia that is caused by the deforestation in the tropics is shown in the following figure.
From a dynamical perspective, it is possible that the tropical deforestation may also induce a Northern Annular Mode (NAM) - like oscillation. The following figure shows the change due to deforestation in the 200 hPa geopotential height field for January. Note the strong reduction in heights over the polar core and the increase in heights in the extratropics. This pattern is characteristic of a positive mode of the NAM. While it is unlikely that tropical deforestation could induce a switch in the mode of the NAM, it is possible that it could contribute to strengthening or weakening the mode and having an impact on Northern Hemisphere climate change.
It is possible that all three tropical forest regions are contributing to the warming observed in the model, however, the following figure suggests that the warming is mainly due to the Amazonian deforestation and that the other regions are secondary contributors that may only amplify the warming by a small amount. This result is not all that surprising when one considers that a logical response to the deforestation is likely to be "felt" poleward and downstream from the forcing region. In this case, northward and eastward to Eurasia. Other studies have suggested that the climate response to Amazonian deforestation would likely be detected in North America, but from a climate dynamics perspective, this does not entirely make sense nor do these model results suggest it. Finally, results from deforesting the other tropical forest centers show that Africa has only a weak extratropical influence and the Indonesian Archipelago has none at all.
Currently, our group is examining how more realistic (fractional) deforestation might affect these teleconnection processes. Preliminary results suggest that there is a linear climate response up to a threshold, at which point the teleconnection response becomes highly nonlinear and noise dominates.
Although this study is highly idealized - the tropics are unlikely to undergo such widespread deforestation (hopefully). However, we should be thinking about land use and land cover change-induced teleconnection processes. While the global teleconnection processes are likely not as important as the regional teleconnection processes, they do show that the potential exists and that a significant portion of the climate change signal may be due to land use and land cover change, not just the radiative effects from increases in atmospheric greenhouse gases.