10.4.1. Feedbacks and Interactions
Several integrated studies have looked at feedbacks and complex interactions
in African regional systems. In the west African Sahel, land surface-atmosphere
interactions have been examined in great detail to explore their role in interannual
variability of rainfall since the long drought that started in the late 1960s.
Reviews by Nicholson (2000) and Hunt (2000) summarize the state of knoweldge
for the physical climate of the Sahel. In general, surface processes modulate
rainfall variability, along with SSTbut in complex ways. The Sahel is
likely to remain a major study topic, and field-based observational studies
are on the rise. A field campaign in 1992 called the Hydrologic Atmosphere Pilot
Experiment (HAPEX)-Sahel was designed to find ways of improving modeling
of land surface properties; a series of HAPEX-Sahel papers were published in
a special issue of the Journal of Hydrology (Goutourbe et al.,
1997). More recently, a regional land-atmosphere experiment is underway in southern
Africa to study fires and emissions, with modeling studies planned to explore
land-atmosphere linkages (see <http://safari.gecp.virginia.edu>,
accessed October 2000). These studiesand integrated modeling in generaloffer
comprehensive tools for studying the integrated Earth system. More studies of
this kind will assist in our understanding of linkages between land surface
processes, regional and global linkages, and human activities.
Biomass burning plays an important role in global atmospheric chemistry (Andreae,
1991), particularly with respect to generation of trace gases that lead to the
formation of tropospheric O3, carbon monoxide (CO), nitrogen oxides,
CH4, volatile organic carbon (VOC, which also is a GHG), and smoke
particles (aerosols), which have an anti-greenhouse effect. Africa is a significant
location of biomass burning. In Africa, there are three main types of biomass
fires: those associated with land clearing for agriculture, which are mostly
located in humid tropical forests and the subhumid tropics; burning of wood
for domestic energy (either directly or after first converting it to charcoal);
and fires in natural and semi-natural vegetation, which are not associated with
changes in land cover or use. Emissions from all three types are of broadly
comparable magnitude (Scholes and Scholes, 1998), although the last category
has received the most attention.
Fires in natural vegetation are not considered to be a net source or sink in
the global carbon cycle because when integrated over large regions and over
several years, CO2 (as well as CO, CH4, and VOC, which
ultimately converts to CO2) emitted by the fire is taken up again
by vegetation regrowth. This is true if the overall fire frequency or intensity
is not changing, but if fires become more frequent or consume more fuel over
time, the result will be a net CO2 source; conversely, if fire frequency
is reduced or the fires burn less fuel, a carbon sink will result, manifested
as an increase in woody biomass. There is no evidence that at a contental scale,
fires in natural vegetation have increased or decreased in frequency or intensity
in the historic period. For some subregions (such as parts of southern and east
Africa), where there is clear thickening of the woody vegetation, it is likely
that fire regimes have become less frequent and intense during the 20th century.
During years of regional drought, such as those in southern Africa associated
with El Niño events, the area burned decreases by about half (Justice
et al., 1996). It is believed that this is caused principally by a decrease
in fuel availability.
The potential for teleconnections in impacts of land-use change on distant
climates further increases the risk in communities that may be at low risk but
will be impacted by actions taken in distant areas. For example, deforestation
of the central African basin leads to climatic impacts in the savannas to the
south in GCM modeling experiments (Semazzi and Song, 2001).
It is clear that rainfall (e.g., intensity) combined with land-use conversion
in watershed areas leads to increased soil erosion. Enhanced siltation in rivers
and increased use of chemicals also leaching into the runoff interferes with
river chemistry (e.g., eutrophication), with major implications for water quality
in lakes and coastal systems. Impacts on biodiversityhence important economic
fisheries, and consequent feedback on national economiesis an area of
research that needs elucidation.
10.4.2. Uncertainties and Risks
There is great uncertainty about how climate might change at subregional scales
in parts of Africa, especially how this might be influenced by human-driven
factors such as deforestation and alternative land uses. Regional climate modeling
will help reduce these uncertainties, and there are early results now of such
modeling for parts of Africa. For these results to be useful, they will have
to incorporate realistic disturbance regimes (e.g., realistic deforestation
mechanisms and land-use characterizations).
Climate change will manifest itself through changes in climate variability
and hence changes in extreme events. Given its socioeconomic status, Africa
is unlikely to respond any better to extreme events than it has in the past.
Flooding, droughts, and so forth are increasingly difficult for Africa to cope
with given increasing pressures on resources from rapid population growth and
dwindling resources. Most African countries remain largely unable to gather
adequate data and information in a timely manner to address critical problems
and surprises such as droughts and floods. Although progress is being made to
design environmental information systems, models for analyzing impacts and policy
options are largely nonexistent. Although adequate data probably exist, what
is needed is the capacity to access large amounts of data and synthesize it
into useful bits of information for decisionmaking. For example, satellite data
have been collected over the past 2 to 3 decades, yet their use is largely restricted
to mapping and short-term climate predictions. Effective information systems
and monitoring have not been achieved.
Evaluation of impacts in monetary or other quantitative terms remains a major
obstacle to comprehensive assessment of impacts of climate change for Africa.
Regional integrated assessment modeling, such as in Egypt (Yates and Strzepek,
1998), offers a solution, and model development should be accelerated for subregions
of Africa where building blocks exist (Desanker et al., 2001).
ENSO-related impacts remain uncertain, given perceived changes in ENSO events
in terms of frequency and duration. However, much progress is being reported
in prediction of ENSO, and this information should be closely linked with case
studies of how different regions and populations respond to specific climate-related
events. These studies should document costs and benefits, as well as responses.