Assessment of the Impact of Land Use Change on Climate – Phase I

Changes in land use alter the exchange of energy and water between the atmosphere and the earth’s surface. For example, the conversion of forest into agricultural land changes the reflectivity and roughness of the surface, as well as its capacity to retain and release water through evapotranspiration. These biophysical effects have a significant influence on climate at regional and global scales. However, regional climate models respond very differently to such land use changes.

The main aim of this project is to assess the response of the CRCM5 and the Canadian Land Surface Scheme (CLASS) to severe changes in land use, and to compare this response to other regional models used by the international scientific community. This is an essential preliminary step toward evaluating the impact of land use on historical and future climates during phase II of the project. The simulations produced will also be used in a related project to assess the impact of land use on flooding.

• Assess the sensitivity of the Canadian Regional Climate Model, Version 5 (CRCM5) and the Canadian Land Surface Scheme (CLASS) to a severe land use change, namely complete afforestation or deforestation of non-desert regions;

• Quantify the biophysical effects of these land use changes in North America and Europe by comparing the responses of various regional climate models and land surface models;

• Share the data generated in order to assess the impact of land use on flooding.

• Produce ERA-Interim driven CRCM5 simulations over two domains, North America and Europe, with two vegetation covers: fully forested (Forest) and deforested (Grass), following the Land-Use and Climate Across Scales (LUCAS) protocol;

• Compare the climatology of the Forest and Grass simulations to assess the sensitivity of CRCM5 and CLASS to this severe land use change;

• Compare the results of the CRCM5/CLASS simulations over the European domain with the ensemble of simulations produced by LUCAS;

• Initiate a concerted comparison of various regional climate models and land surface models as was done for LUCAS, but for the North American domain;

• Compare the set of simulations produced over the North American domain with the LUCAS ensemble for Europe to assess the transferability of the results.

Production of a new simulation ensemble

An ensemble of six climate simulations was produced with the CRCM5 to test the sensitivity of the model to major changes in vegetation: the complete deforestation or afforestation of North America and Europe. Since the LUCAS protocol was followed, it was possible to compare the sensitivity of the CRCM5 in Europe with the LUCAS ensemble, which includes nine combinations of regional models and surface models. Researchers from the National Center for Atmospheric Research and UQAM also participated in the initiative by producing simulations with their own models on North America. Thus, not only was the CRCM5 response analyzed in depth, but it was possible to compare this response to an ensemble of simulations on both continents. 

Winter response

Like most climate models, the CRCM5 simulates significant biophysical warming (i.e. excluding the effect of greenhouse gases) due to the presence of forests. This warming is caused by a decrease in the albedo of the surface. In the Herbaceous simulations, vegetation is easily buried in snow (high albedo). Conversely, in Forest simulations, the dark canopy is made up of coniferous trees, which mostly retain their needles during the cold season and can thus hide the snow on the ground (low albedo). Thus, a greater fraction of solar energy is absorbed by the surface, which causes warming that is mainly concentrated in boreal forests. This is more significant in North America than in Europe, because coniferous forests extend to lower (and therefore sunnier) latitudes in North America. Compared to the other models in the LUCAS Europe ensemble, the CRCM5 has a response with a similar spatial pattern, but with the most pronounced amplitude. The WRF Noah model produces an unreliable response because the snow hiding mechanism is not possible. Thus, its response is weak compared to the two versions of the CRCM, which are very similar.

Figure 1 : Winter response of the CRCM5 to afforestation (Forest-Herboceous simulations) averaged over the years 1986 to 2015.

Summer response

The signal is much more complex in summer, and the models struggle to agree even on the sign of the temperature response. The source of this disagreement seems to lie at the heart of the surface models, which produce different partitions between sensible and latent heat flows. This discrepancy was also noted in an ensemble of global models subject to a similar sensitivity experiment. Vegetation strongly influences the effectiveness of surface evapotranspiration. For example, a deciduous forest is typically associated with greater transpiration than herbaceous land (this is the case in CLASS, but not in all surface models). As a general rule, the more a category of vegetation promotes latent heat flows (evapotranspiration), the more sensible heat flows will be reduced. This will result in relative cooling, as energy is invested in breaking hydrogen bonds rather than heating air near the surface. This is indeed what is observed in simulation ensembles in North America and Europe: models that demonstrate an increase in latent heat flows at the expense of sensible heat flows tend to produce summer cooling, and vice versa. Both versions of the CRCM/CLASS produce cooling over deciduous forests. The effect on evapotranspiration of replacing boreal forests with herbaceous cover is less clear. The decrease in albedo then dominates, and the dark canopy of coniferous forests causes summer warming at high latitudes. Thus, the summer temperature response of CRCM5 forms a dipole, with warming at high latitudes (coniferous forests) and cooling at low latitudes (deciduous forests). The average of the summer responses of the models in the LUCAS ensemble also produces such a dipole, but of lower amplitude than that produced by the CRCM5.

Summary

The Forest–Herbaceous experiment made it possible to evaluate the sensitivity of CRCM5-CLASS to drastic changes in land use, and to compare this sensitivity to that of a multitude of other regional climate models and surface models. Analysis of these simulations confirms that the CRCM5-CLASS is favourably situated in relation to the ensemble, in that it seems to adequately represent important processes (changes in albedo due to snow and the efficiency of evapotranspiration, for example) and its response is comparable to that of the rest of the ensemble. This new knowledge will be essential in carrying out the second phase of this project, which aims to assess the climate impact of realistic land use changes.
 

• Université du Québec à Montréal

• NCAR

• GERICS

• CCCma

• Direction de l'expertise hydrique (MELCC)

• École de technologie supérieure

• LMU

• LRZ

706800