and I know it, because news stories like this one about the flooding in Niger hit me a completely different way now – previously, I would have thought about how this could be teachable, and even how it might relate to some research ideas . . . now, I recall interviews from April with people in my new Bureau at USAID where we discussed the looming food crisis in Niger.  In mid-September, this won’t be a teachable moment – this will be a fire drill for which I have some degree of responsibility.  Sobering.

Incidentally, this is another example of the challenges that face those of us working at the intersection of environment and development.  The long-term (last four-five decades) signal for precipitation is in steady decline.  It is hard to say if this is a visible outcome of climate change, mostly because we have a lot of trouble understanding the mechanics of the West African climate (for those so inclined, there are some issues with the teleconnections from ENSO and the influence of the NAO).

Dunkwa (Ghana) weather station precipitation figures 1963-2000 (source: Ghana Meteorological Service)

This figure (from my upcoming book) illustrates the real problem, though – the long-term decline is clear at this weather station (the closest one to my research area that is not parked right on the beach), but more striking is the variability around the centerline.  While this station is not showing any real trend toward greater variability, many other places in West Africa are – hence the massive, surprising flooding we are seeing in Niger, despite a long-term trend toward less precipitation in the region.  People forget that there are two key variables that shape precipitation outcomes – amount and timing.

This is probably the hardest part of the job – thinking about how to plan for increasing unpredictability and variability.  Trends are easy, assuming their mechanics are understood and therefore somewhat predictable.  If I know there will be 10% less rainfall in a particular place by a particular year, I can go about figuring out what the biophysical, economic and social impacts of that change might be.  However, it is a hell of a lot harder to plan for 10% more variability by a given year (assuming we could even quantify rising variability in such a manner).  Well, if it was easy, it wouldn’t be interesting . . . and someone else would have solved it already.