usaid/nambia environmental threats and opportunities assessment
usaid/nambia environmental threats and opportunities assessment
usaid/nambia environmental threats and opportunities assessment
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
ANNEX C<br />
Interpretation of CC projections for Namibia (After Turpie et al 2010)<br />
Because of the high level of uncertainty in projections for future climate in southern Africa, particularly<br />
of rainfall change, it seems useful from a policy perspective to estimate the potential ranges of impacts,<br />
including high, median <strong>and</strong> low impacts, by 2050, as it is unlikely that the uncertainty range will be<br />
reduced in the near future, <strong>and</strong> because of the impacts of current levels of climate variability in the<br />
region. Variability is likely to dominate the climate signal for at least a few decades until clear climate<br />
change signals become evident. Using high resolution spatially downscaled climate information seems of<br />
little use in this regard, as it is more important for policy development to estimate the impacts<br />
particularly at the median <strong>and</strong> “tails” of the distribution of possible future climate scenarios. Estimates at<br />
the tails of the distribution can provide an <strong>assessment</strong> of impacts that have a low probability but a high<br />
societal relevance if they do occur.<br />
Climate scenarios that are currently generated using General Circulation Models (GCMs) have two main<br />
sources of uncertainty that result in a relatively wide range of projections, especially for rainfall futures,<br />
for southern Africa in particular. These are 1) the GCM design itself, which varies between the several<br />
models used in the IPCC AR4, <strong>and</strong> 2) the emissions scenarios used to drive the GCMs. The largest<br />
source of uncertainty by the middle of this century is due to GCM design, <strong>and</strong> rather little is due to<br />
emissions scenario. Emissions scenario is however an important source of uncertainty <strong>and</strong> variation for<br />
simulations towards the end of the century. As mentioned above, due to the potentially large range of<br />
uncertainty in scenarios, it seems of little value to focus on fine spatial scales for climate scenarios <strong>and</strong><br />
impacts studies, as by far the largest source of uncertainty is at large spatial scale. It is also of limited<br />
value to consider a range of emissions scenarios, but rather to focus on underst<strong>and</strong>ing the range of GCM<br />
variation, <strong>and</strong> to attempt to represent impacts that might relate to the median <strong>and</strong> the extremes of that<br />
range for policy relevant information.<br />
Unfortunately it is currently difficult to obtain spatially downscaled climate projection data for measures<br />
other than rainfall or temperature for southern Africa outside of South Africa for the IPCC AR4 climate<br />
projections, especially for the middle of this century. We have thus compared the best available<br />
information for the IPCC AR4 generated by GCM’s for the year 2100 (median projections of 21 GCMs,<br />
driven by the A1B emissions scenario) with the interpolated HADCM3 GCM data used for the previous<br />
most comprehensive impacts <strong>assessment</strong> on Namibia for 2050 ( Midgley et al 2005 . Because this<br />
comparison (Figure A) shows that the HADCM3 GCM used by Midgley et al. (2005) represents roughly<br />
a median climate future for the 21 AR4 GCMs, climate surfaces representing rainfall <strong>and</strong> temperature<br />
change at the monthly temporal scale for 2050 <strong>and</strong> 2080 have been created for this project using the<br />
HADCM3 GCM (as driven by the A2 scenario). These have been overlaid on a current climate surface<br />
that is taken from the recognized <strong>and</strong> quality-controlled WorldClim data set <strong>and</strong> used for impact<br />
<strong>assessment</strong>s of species-level change.<br />
Comparison of IPCC AR4 scenarios with those used by Midgley et al (2005) reveal that the median<br />
rainfall change projected for 2100 by the IPCC AR4 (between 5 <strong>and</strong> 20% reduction) is comparable to the<br />
least extreme median rainfall change used by Midgley et al. (2005), represented by the HAD CM3 model<br />
for 2050, under an A2 emissions scenario. By 2080, this scenario suggests a more extreme rainfall change<br />
of between a 10 <strong>and</strong> 30% reduction. The 2050 scenario used by Midgley et al. (2005) shows a relatively<br />
spatially uniform rainfall change, with the largest reductions of ~ 20% across the centre of Namibia, with<br />
more severe drying suggested in the northwest <strong>and</strong> on the central coast. This contrasts with the IPCC<br />
USAID/NAMIBIA ENVIRONMENTAL THREATS AND OPPORTUNITIES ASSESSMENT 95