Previous work using climate models found that a hypothetical replacement of the boreal forest in Canada and Russia with bare land increased the surface albedo from 0.1 to 0.5 at near latitude 60?N as snow cover replaced the tree cover (122, 123). This resulted in significant cooling throughout the year of 2-3?C at areas as far south as 30?N with a 1?C decrease in air temperature predicted for areas near the equator. During the spring, as much as a 12?C decrease in April was predicted for areas around 60?N. This was partly a result of there being more land on which snow could accumulate and reflect sunlight. Changes on this scale might be disastrous. However, this was based on an instantaneous change in surface albedo averaged over a 5-15 year time period, whereas we are proposing incremental changes so that the coverage could be stopped or reversed if necessary.
Other attempts at modeling land use change have also shown that changing the surface albedo causes significant temperature increases or decreases in the affected areas (124). A noteworthy difference in our plan is that it will be implemented incrementally over 60 years. The previous modeling studies are based on instantaneous hypothetical changes in land albedo followed for decades. Thus, while the types of temperature cooling found in the boreal forest modeling might still occur, there would be more time for the colder areas created to equalize with the rest of the globe.
Most of the modeling done involving the deserts has concentrated on the impact of droughts on expansion of the Sahara into the bordering area known as the Sahel. The general conclusion is that an increase in the albedo of the southern Sahara would cause cooling of the Sahara, less evaporation of water and lead to more subsidence heating, in which dry air from the Sahara spills over into the Sahel, drying it out further and expanding the desert (17). These models are far from definitive and the lack of surface monitoring stations in the Sahara as well as other deserts greatly limits the quality of input data for them.
Although an impervious cover will limit evaporation of water that gets underneath it, the cover will also facilitate evaporation of water on the surface. Thus, more water will likely go into the atmosphere than would be the case if bare ground were present. This in turn may mitigate the increased subsidence heating of the Sahel. Such a scenario should be considered in any modeling of future climate.
In addition to temperature and precipitation related effects of surface albedo change, another area that needs to be studied by modeling is the impact on dust flow patterns. Atmospheric dust from deserts and other areas contribute to reducing overall radiative forcing by reflecting sunlight. Some of this dust also serves to fertilize the phytoplankton in the N. Atlantic and the forests of the Amazon. The possible impact on a reduction of this dust loading to either of these must be studied before large-scale surface changes are begun. It may be possible that the modeling will identify critical areas that can be covered without impacting dust loading and flow patterns as well as those that cannot. Similar concerns have been expressed about strategies to increase carbon sequestration in deserts through increased growth of drought tolerant vegetation (75).
A recently advanced theory that land use changes and specifically albedo changes have a greater impact on climate than radiative heating due to GHGs could also be evaluated in the modeling to be performed. This work suggests that albedo effects cause local and regional changes in surface heat distribution that change climate more than would be predicted by simple radiative heating (107, 108, 125, 126). One concern is that afforestation, covering previously bare or higher albedo land with new forests to reduce atmospheric carbon dioxide levels, might also produce disruptive climate change. This is most likely to occur when trees are planted in areas covered by snow much of the year like the tundra (126).
The specific locations to be modeled before and after albedo enhancement have not been identified yet. We expect that the Sahara, the Arabian and the Gobi deserts will be included because of their large potential coverage area. Modeling the impacts on covering very dry low albedo deserts like the Atacama would also be useful in predicting future outcomes.
Finally, if a significant test area is covered, e.g., 10-1000 square miles in the Mojave or Sahara, it may be possible to compare the effects on local weather using weather forecasting models vs. observed conditions to see if the coverage changes the weather.
