 |
||||
|
||||
|
|
 |
||
|
||||
      |
Climate and Land Use Change Processes in East Africa
This project is no longer current. Please see the Research Programs page for a list of current research projects.
While some regions of East Africa are being preserved as natural areas, others, including the lower slopes of Mt. Kilimanjaro, are being converted to agriculture. Some members of the Maasai tribe, traditionally a pastoral people, are changing to an agricultural way of life.
Collaborators
Jeffrey Andresen-Dept. of Geography, Michigan State University (MSU
web site)
David Campbell-Dept. of Geography, Michigan State University
Ruth Doherty-Dept. of Physical Geography, University of Edinburgh, UK
(Univ. of Edinburg web site)
Marianne Huebner, Department of Statistics, Michigan State University
David Lusch-Remote Sensing Center, Michigan State University
Stephen Magezi-Ugandan Meteorological Service and Makarere University
(Makarere Univ. web site)
Joseph Maitima-International Livestock Research Institute, Nairobi, Kenya
(ILRI web site)
Nathan Moore, Department of Geography, Michigan State University
John Nganga-Dept. of Atmospheric Science, University of Nairobi (Univ.
of Nairobi web site)
Jennifer Olson-Dept. of Geography, Michigan State University
Jiaguo Qi-Remote Sensing Center, Michigan State University
Jean Palutikof-Climate Research Institute, University of East Anglia (Univ.
of East Anglia web site)
Bryan Pijanowski-Dept. of Zoology, Purdue University (Purdue
Univ. web site)
Robin Reid-International Livestock Research Institute, Nairobi, Kenya
Fredrick Semazzi-Marine, Earth, & Atmos. Sciences, North Carolina State
Univ. (NC State web site)
Philip Thornton-International Livestock Research Institute, Nairobi, Kenya
(Univ. of Nairobi web site)
Pius Yanda-Dept. of Geography, University of Dar Es Salaam (USDM
web site)
Executive Summary
Brent
Lofgren of GLERL has been collaborating with a large research group centered
at Michigan State University in the Climate-Land Interaction Project (CLIP web site). The overall objective of the project is to establish the mechanisms
that determine how climate and land use (primarily in the human-mediated
realm, but also land use change through natural processes) interact with
each other and possibly establish a feedback between the two. The work
of GLERL researchers is primarily in the climate modeling phase of this
project. Input from GLERL has also been very important for the crucial
planning of experimental design. The work described here involves climate
modeling with various prescribed states of land use, and no direct linkage
yet to the other phases of the project, which aim to predict, on the basis
of climate, such things as crop yields, land tenancy of farmers and pastoralists,
urban expansion and transformation. However, we have used input from the
section of the project that uses remote sensing to translate land use
types into physical characteristics. The ultimate goal of the project
is to "close the loop," incorporating all of the feedback between the
climate and human parts of the system, but this has not been achieved
yet. The work described below, however, demonstrates a section of this
loop and the sensitivity of climate to land use, a prerequisite for the
potential of feedback.
Rationale
This project will address how the physical characteristics of land surfaces are likely to interact and feed back with climate processes. As illustrated in Fig. 1, the research activities will center on different aspects of the land-atmosphere system, with connections between different sub-teams of scientists and their respective sets of data and predictions. The whole of this system may include feedbacks, by which climate can effect net primary productivity (NPP), which can affect land use and land cover, and in turn climate again. (Note on definitions: "Land use" refers to the human use of the land or lack of use, while "land cover" means the physical characteristics of the resultant land surface.)
The climate modeling box in the figure will be filled by a version of our CHARM (Coupled Hydrosphere-Atmosphere Research Model) and possibly RegCM2, a similar model. The NPP simulations box will include models of the response of both natural and agricultural plants to changes in climate. The land use box will involve further modeling of the effects of these concurrent changes in climate and NPP, and the land cover box will involve a mapping of the land use results to physical characteristics of the surface.
Methods
A series of sensitivity tests were undertaken to optimize specific parameters in the Regional Atmospheric Modeling System (RAMS).
- Cloud condensation nuclei (CCN) concentration
- CCN shape parameter
- convective updating frequency
- Convective trigger velocity
- radiation scheme
- radiation update frequency
- Initial soil moisture conditions
Our primary research activities have been to conclude the validation and to develop precipitation comparisons of the model to observations. Following this, our activities have centered on incorporating an accurate representation of leaf area index (LAI) and fractional cover, based on satellite-derived quantities, and evaluating the model's response to this improved parameterization.
Atmospheric observations in East Africa are scarce. Validation of the model was conducted based on comparisons with radiosonde data where available, and with remotely-sensed data if applicable. Since the CLIP project places emphasis on integration of many parts, we have emphasized validation of the model against precipitation and temperature. These two variables will be relayed to a crop model, so it is imperative that RAMS output be validated against precipitation and temperature at a minimum. These variables are governed in large part by the radiation budget at the surface, so accurate spatial, temporal, and phenological representations of the surface are vital to accuracy.
Upon completion of validation, we focused our research activities on comparing the default Olson Global Ecosystem (OGE) land cover with a new land cover hybrid, which we call CLIPcover. The first stage of this comparison was to replace the spatial distribution of OGE land cover classes with the CLIP cover distribution. The phenological and temporal variability of these land cover classes was not altered at this point. After 1 month of simulations using both land cover schemes, the root-mean-squared differences in accumulated precipitation compared to Tropical Rainfall Monitoring Mission (TRMM) estimates are statistically indistinct. Spatially, modeled rainfall reproduced the inter-tropical convergence zone cloud cover but generated much more precipitation at higher elevations. The levels of increased precipitation appear to be related to changes in albedo and shifts in large-scale transport of moisture (Figure 1). Albedo is not well-correlated with precipitation in the southernmost part of the domain and in the Lake Victoria region. These anomalies may be related to differences in upper boundary layer winds and lake temperature respectively, but this is still under investigation.
 |
 |
Figure 1a. Simulated accumulated
precipitation (mm), March 8-31, 2000 (CLIP cover minus OGE cover) |
Figure
1b.Simulated albedo difference, March 8-31, 2000 (CLIPcover
minus OGE cover) |
Integration with Land Cover
In most atmospheric models, land cover phenology is represented simply as a function of latitude and Julian day; this is the case with RAMS. However, east Africa is unique among equatorial regions in its low LAI, lack of dense rainforests, and bimodal rainfall pattern. This sharp departure from typical phenology necessitated an improved representation of land cover and a more accurate depiction of vegetation properties-namely, LAI and fractional cover- over time. The LAI splines constructed by Lijian Yang and Jing Wang clearly capture the bimodal character of east African vegetation, particularly for maize farming, and these splines have been incorporated into RAMS. We anticipate reproducing this spline approximation for fractional cover as well in RAMS. Figure 2 shows the changes in LAI resulting from the change to CLIPcover, followed by the addition of the LAI spline function. The Moderate Resolution Imaging Spectroradiometer (MODIS) image for that same date is given for comparison. Errors in classification still exist, particularly in the southern parts of the domain, but the overall representation of LAI in the model is improved.
Figure 2. LAI for east Africa using OGE cover, CLIPcover, and CLIPCover plus LAI spline approximation. MODIS imagery for the same date is provided for comparison. Pale blue in the MODIS image represents water or cloud cover.
Products
Ge, J., J. Qi, N. Moore, N. Torbick, B.Lofgren, and J.M. Olson, under review, Impacts of Land Use/Cover Classification Accuracy on Regional Climate Simulations, J. Geophys. Res.-Atmospheres.
Lofgren, B. M., N. J. Moore, J. A. Andresen, J. J. Olson, D. J. Campbell, and B. C. Pijanowski, 2005: CLIP: Climate-Land Interaction Project-Investigating human-climate interactions in East Africa. 15th Symposium on Global Change and Climate Variation, 85th Annual Meeting, American Meteorological Society, San Diego, CA, 9-14 January 2005. Available on CD-ROM.
Olson, J.M., G. Alagarswamy, J. Andresen, D.J. Campbell, J. Ge, M. Huebner, B.Lofgren, D. P. Lusch, N. Moore, B.C. Pijanowski, J. Qi, N. Torbick, and J. Wang. under review. Integrating Diverse Methods to Understand Climate-Land Interactions at Multiple Spatial and Temporal Scales, GeoForum.
Presentations
Andresen, J., B. Lofgren, R. Doherty, F. Semazzi, and N. Moore (presented by J. Andresen), 2004. Climatological research activities associated with the CLIP project: Regional climate modeling. American Association of Geographers Meeting, Philadelphia, March 15-18, 2004.
CLIP team (presented by Campbell and Pijanowski). 2004. Modeling Land Use Change and Precipitation in East Africa - The CLIP Project. NASA ESSN Symposium, Water and Energy Session, Sept 27-29, 2004, College Park, Maryland.
Lofgren, B. M., N. J. Moore, J. A. Andresen, J. M. Olson, D. J. Campbell, and B. C. Pijanowski, 2005. CLIP: Climate-Land Interaction Project-Investigating human-climate interactions in East Africa. 15th Symposium on Global Change and Climate Variation, 85th Annual Meeting, American Meteorological Society, San Diego, CA, 9-14 January 2005.
Moore, N. J., B. M. Lofgren, and J. A. Andresen, 2005 (presented by Lofgren): CLIP: Modeling Land Use Change and Precipitation in Eastern Africa. 19th Conference on Hydrology, 85th Annual Meeting, American Meteorological Society, San Diego, CA, -14 January 2005.
Pijanowski, B., M. Huebner, B. Lofgren (presented by Lofgren and Pijanowski), 2005. Sources, Characterization, and Communication of Uncertainty in Climate Impacts, Workshop on Climate Science in Support of Decision Making, U.S. Climate Change Science Program, Arlington, VA, 14-16 November 2005.
