I’ve been quiet for a while, despite my promises to do better. At least I have some sort of excuse, which is that I spent the last few weeks in Thailand to work with Wirong Chanthorn from Kasetsart University in Bangkok on the analysis of a dataset on successional dynamics of a tropical forest that he collected in the last year.
The forest data is from a number of medium-size (0.5 ha) plots located in Khao Yai National Park. I’ve been there only for a short time (most of the time we were in Bangkok), but Khao Yai seems to be a great place to do research: accessible, comparatively save, climatically quite agreeable and civilization not too far away when needed. Probably there are more wild places, undisturbed from jeeps taking tourists to cruise around for wildlife at night, but great to get some work done without having to worry about running into an illegal drug plantation or worse, as one might have to in some other parts of the world.
I had first heard of Khao Yai because of the large Mo Singto 30 ha forest dynamics plot maintained by Warren Brockelman, which is part of the CTFS plot network. Outside the “tropical forest community” though, Khao Yai may be better known for research on animals such as the charismatic hornbills as well as apes and gibbons in particular. In fact, Warren Brockelman, who went with us to visit the plots, started out working on gibbons in Khao Yai before establishing the Mo Singto plot, and gibbons remain one of his main interests.
As most of the CTFS plots, Mo Singto is located in an area covered by old-growth forest (whether it’s really primary or rather old secondary forest is not completely clear). But not all of Khao Yai’s forests are late-successional. Apparently, the first settlements of local villagers in the area appeared in the 1920s. Due to its remoteness, the region subsequently turned into a refuge for fugitives of all sorts, who were hiding in inaccessible areas of the forest. These activities resulted in smaller and larger disturbances all over the area, although old-growth forest remained in most of the area. Human impacts decreased strongly when the national park was established in the 1960s and the villagers were relocated from the area.
Due to this history, there is quite a bit of secondary forest of different ages in the area, as well as some grasslands that are kept open by prescribed burning from the National Park. I guess the latter is mostly to provide better visibility of wildlife such as elephants and sambar deer, which are abundant in the forest. This mix of different vegetation stages makes the park an interesting place for studying forest succession, as well as an interesting area for remote sensing studies.
For our project, the secondary forest areas were the main focus. To be more precise, we were looking at the differences in stand structure through succession. We’re not quite through with the data analysis, but I guess it’s fair to say that, as one would expect, there are some pronounced differences in stand structure throughout the different successional stages. I’m sure I’ll report about the results when they are in a more mature stage.
Working with the data and speaking with Warren and Wirong about the large amount of work and dedication it needed to create them coincided with a cascade of blog posts on data sharing, triggered by a change of PLOS’s data sharing policy (required to deposit raw data). I have made a few comments to some of those posts, and I do have vague plans for a more general post on data sharing in the future. For the moment, suffice to say that working on the data was a good reason to reflect about the ethics of ownership, but also about the work needed to extract knowledge from ecological data. An example is the seemingly indefinite number of analyses that turned out to be possible with the “mother of all CTFS plots”, Barro Colorado Island.
It’s a huge investment to create these type of data (I think the first census of the Mo Singto plot was about 11 men-years), and the people that did this work certainly have some ethical right to have priority for using the data. Yet, a lifetime may not be enough to do all the analyses possible with these datasets, and for most of the CTFS (apart for BCI possibly), more time has been spent on data creation than on analysis. It seems to me that, to make the most of these great initiatives, it seems key to find ways for efficient data sharing and analysis with proper attribution, but also with minimal barriers for the people that want to be working on the data.