Monday, June 21, 2010

Ecozone Prediction From First Principles

We've all had that experience of running or hiking somewhere and saying, "You know, right here in this valley/mountain/prairie, you could trick me into thinking that I was in x." You might be thinking of another spot on the West Coast of North America, or you could be thinking of somewhere else entirely. As a recent transplant from NorCal to SoCal, I'm always noticing when somewhere in SoCal reminds me of NorCal. In the winter, in Miriam Bear if you squint, you can almost pretend you're in some parts of Marin County, and some of the parks in the hills behind Malibu similarly remind me of some of the more rugged East Bay parks.

I also find myself guessing more specifically at the impact to San Diego of various hypothetical climactic changes; instead of alternative history, it's alternative climatology. What if that wet winter San Diego had extended into June? What if this happened every year from now on? Unfortunate though it may be, we wouldn't look like British Columbia overnight; there aren't any Douglas fir cones around, for one thing, and the ferns would have to spread out of the very few canyons where we can now find them around here.* It would take awhile; after all, after the glaciers melted off of Europe 12,000 years ago, the preboreal period in Russia lasted a good seven centuries.

Of course such geeky speculation led me to further wonder whether you could take climate and geographic data and inductively fit an equation that would predict resulting vegetation with any useful kind of accuracy. For the ecozone values you could try to match your outputs to you could use the Köppen system. The most useful inputs would be:

Latitude - the further south you go, the more desert like it is. Pretty obvious, but it doesn't explain why there are redwoods in SLO County and deserts in British Columbia. The transition is often abrupt. Once in inland Washington State, I came out of the evergreens into the glaring July sun of a Western desert in the space of less than five miles - it was that abrupt (on Route 97 at the southern edge of Wenatchee National Forest.)

Distance from coast - again obviously, the further inland you are, the drier it is. Taking only the first two factors, a simplifed North America, e.g. completely flat, would look like the more gay-friendly map on the left, but in fact it looks like the map on the right:



Right about now you're wondering not without justification whether I've completely lost my mind. But do notice that I used humor just right up there. Because humans have been demonstrated to react positively to it. Also note that the no-mountains map isn't completely insane relative to the real-world map.

Of course the complication to the gay-friendly map is the terrain - elevation, drainages, plus angle and consistent gain of ranges. This is the tricky part of the equation, and it makes all the difference. A 30 degree slope for 100' of elevation gain? Who cares. For 3,000' of gain? If it's west-facing, it'll be greener. Look at the map of rainfall in San Diego County and you'll see what I mean. You can clearly see in the state map above the Cascades, plus the coast ranges and the Sierras. If you're in the lee of a real slope, you'll be drier and have less moderating benefit from the ocean. This is why Santiago, Chile has more extreme weather than Sacramento or even Bakersfield- Chile's coast ranges (as opposed to the next range inland, the Andes) attain 10,000'+, as opposed to California's which top out at ~6,000' with a ridgeline usually closer to 3,000'.

If you want to get really fancy you could use some measure of "evenness" of rainfall, i.e. what's the standard deviation on monthly precipitation? In Mediterranean climates this is high. San Francisco gets the same rainfall as Dodge City Kansas, except it pretty much all falls in a five month period from November through March. In Dodge City it falls pretty evenly across the whole 12 months.

But am I going to do this? No! The real proof in the pudding would be in comparing the real Köppen zones at some decent level of resolution (pixels of 100 sq mi?) and including the elevation information would require programming skills and time that I don't have. Also the end result of all this would be merely an exercise in predicting ecological regions from first principles. On the other hand, if there's an Earth-like planet among the 700 new exoplanet candidates that Kepler just unloaded, and you want to know where to buy real estate that's most like Santa Barbara, then give me a call and we can work something out.

*In other news the redwood cones I planted in January in case of massive climate shift haven't sprouted yet.

No comments: