The science of correlation

Ecologists are obsessed with correlative science, urban ecologists even more so. We realize that most natural systems are infinitely complex and cannot be comprehended from a purely mechanistic perspective. One may model protein folding by quantifying the forces of hydrophobic interactions, dipole moments, bond-angle strains, etc., but one cannot exhaustively describe the probability of reproductive success in a generation of Javan Mynahs in Orchard road by a systematic evaluation of every possible (direct) cause to fecundity and mortality. Most natural systems involve too many abstract components to be studied in a thoroughly experimental manner. But that does not mean we surrender to ignorance. Urban ecology (and social sciences in general) emphasizes a pragmatic approach to science that makes logical speculations from generalizable trends. Correlation science is science. …But it is not without problems…and I’ve found myself asking them many times over the last half a sem while studying urban eco.

I cite the recently mentioned example of street trees not as criticism, but in order to incite a more personally-relevant inquiry into the subject. Prof Evans mentioned several times that physiological constraints are the (main?) reason for the high number of exotic trees planted in Singapore. I must qualify before commenting that (i) I do not believe the statement is wrong; the ‘correlation’ does exist (ii) I do not assume Prof speaks presumptuously—I’m most certain he is perfectly well-informed concerning the myriad of other forces that act concurrently and historically to shape the ‘community composition’ of Singapore’s streetscapes. Notwithstanding, the statement is extremely difficult to swallow as it is. Is the fact of “large proportion of exotics in streetscape community” simply explainable by “physiological constraints—Singapore no savannahs, our native trees not adapted”? Many tangents we could pursue. I state a few briefly:

  1. Historic and economic factors. It was only just recently mentioned (was it in a 4265 or 4262 lecture?) that, historically, many plant stocks have simply been more available to streetscape managers than others. Many of these species were discovered and popularized by botanists and arborists who had not been sampling the globe unbiasly (is it coincidence that Hevea brasiliensis, Swietenia macrophylla, and Albizia saman, some of the earliest introduced tree species (1876) are all neotropical?). Certain tree species are simply better studied and “marketed” than others, and very few of these are native to our island for obvious reasons. Tree planting patterns have evolved over the years in Singapore, and most of us have actually observed it in our lifetimes. But the progress of this evolution does not seem to have followed the botanical/horticultural discoveries of more physiologically-adapted tree species as tightly as we would have thought to be the case if physiology were the main/dominant predictor of native/exotic tree community composition. Streetscape managers do shopping too…which brings us to the next factor.
  2. Psychology. It might be discounted anywhere else in the world, but not in Singapore. Because streetscape is managed by relatively few individuals, polarization in “preferences” really make a difference in our streetscape community assemblage. I dare say confidently that Prof Tan from the Botany lab (an individual—individuals should never be “factors” in ecology right?) has probably had more statistically significant influence on the evolution of NParks tree planting philosophy than a good number of other factors we might think of including in any ‘urban tree community model’. And I think many of us can testify to a slight (subconscious?) bias towards foreign imports in this country, which NParks officers might have shared (in the past?) as well. Put another way, the mere idiosyncrasies of an individual streetscape manager ARE going to make a significant difference in our tree community assemblage—how do we harmonize that with the physiological argument? and how do we go about making science out of this?
  3. Lastly, there are actually many native species that grow very well in streets but aren’t commonly planted. Even a dim-witted undergrad like I can name ten or more without looking up a reference book. Singapore was not completely rainforest ecosystem—coastal forests contributed many street tree species and can contribute potentially many more (the SSS1207 textbook discusses this). The Figs are a very good example. Benjamina, microcarpa, kerkhovenii, superba, virens… Thoroughly native, many considered RF spp. They can and do grow well in the urban environment, yet they are not planted much along roads for practical reasons—they damage structure, they have crazy girths, they make scary pillar roots, they shed frustrating volumes of fruits, and their many religious associations can be problematic. Some native tree species which were once extensively used are now being phased out because of practical considerations as well (and perhaps also because we have more money to splurge on other more flamboyant exotic species). Think Cerbera odollam and Terminalia catappa, a 5min thought exercise might reveal why. The argument that Alstonias, Szygiums, Artocarpuses cannot do well in the urban environment is simply groundless because these do grow by road sides in urban environments local and beyond. The fact that exotic savannah tree species offer a wider selection and perhaps additional safeguard against dessication is valid, but one cannot help but realize that the non-biological factors still trump the physiology argument.

What, then, is the point in saying all of this? My question is not about street trees; it is about urban ecology and the science of correlation. The argument from physiology is not wrong at all. Physiological demands place a ‘cut off criterion’ which eliminates quite a few native tree species from the candidature, consequently giving exotic savannah species a much higher chance of being fielded in this game. Yet in the science of infinitely complex systems, explaining a “top 15 trees in cities” histogram with “physiology” may be a slight oversimplification. How far can we extrapolate biologically valid arguments like this one before sounding more like politicians than scientists? Again, I do not propound that Prof Evans is teaching us stuff that is wrong or even inaccurate; and in fact I believe the method of the single take home message is pedagogically powerful. But putting classroom time behind us, is it most rigorously scientific to state the “physiology explanation” without a balanced qualification of the (extremely) strong, confounding (and potentially interacting) effects of the (many) other components of this system?

Or, notated, is

Native/Exotic spp in Tree Community ~ Physiology + etc

a scientifically valid simplification of

N/E.STC ~ HistoricalFactors + EconomicFactors + PracticalConsiderations + IndividualPreferences + Physiology + EF^2 + HF:EF + PC:Phy + EF:Phy + EH:PC + IP:Phy + etc


One thought on “The science of correlation

  • February 26, 2015 at 10:15 am

    Extremely intelligent and well-thought out. Challenging your profs. I like it !!!


Leave a Reply

Your email address will not be published.