Passing Ecological Judgement

Week 3 of Highly Allochthonous’s writing challenge was spent ploughing through very rocky soil in the Methods sections of my papers on Black Hills (South Dakota) high-elevation grasslands.  I was faced with the most challenging part of the project -- assessing grassland quality.  Our ultimate goal is conservation of native montane grasslands in the Black Hills, and towards this end we want to know which are the better and best ones.  But passing judgement on grasslands can be tough and guilt-ridden.

The haygrass timothy dominates many
mountain meadows in the Black Hills.

The most basic criterium for judging these grasslands is “naturalness”, i.e. is a grassland within the natural range of variation?  But here we have a big problem, the subject of an earlier post.  We don’t know much about the structure and composition of Black Hills mountain meadows prior to human settlement, and humans have been busy altering the landscape since the 1880s.  By the 1920s many of the moister meadows already were dominated by non-native hay grasses, as were all of the ones surveyed in 1999 and 2011.  In a way this made the task of evaluation simpler -- any predominantly native grassland would be of interest.

There still are native high-elevation grasslands in the Black Hills, on drier sites that probably were less attractive to homesteaders, sheep and cows.  We ranked these following NatureServe's metric-based ecological integrity assessment framework.  We used three kinds of metrics:  size, landscape context and condition.  The first two were easy.  Size can be measured from imagery in a GIS and there are conventional categories for ranking.  Landscape context was the same for almost all the grasslands we surveyed -- partially disturbed natural or semi-natural vegetation (i.e. national forest managed for multiple use).

Black Hills montane grassland in bottom of Gillette Canyon.  Size = A (>100 ha or 250 ac);

landscape context = B (surrounded by National Forest lands managed for multiple use).  The

unforested uplands in all but the northwest part of the photo burned in the 2000 Jasper Fire.

Assessing grassland condition was the challenge, i.e. assessing in a way that was explicit, consistent, replicable and doable.  I was determined come up with such a protocol, even though the task tied my brain in knots at times.  In earlier surveys “intuitive” rankings had been assigned by “experts” (myself) and I didn’t want to go through that again.  I know that my brain isn't up to the task of consistently evaluating entities as heterogeneous as these large grasslands.

To assess condition we used four metrics:  native plant cover, invasive exotic species (noxious weeds), native increaser species (native plants present in non-natural abundance due to disturbance) and soil surface condition (bare soil or surface disturbance obvious due to altered vegetation).  Each metric could be ranked from A (best) to D (worst).  The categories were fairly broad so that we could assign ranks with confidence.  For example a grassland with <50% native cover was ranked D for that metric.  For a rank of A, native cover had to be >99%.  This sounds simple, but many grasslands were coarsely-heterogeneous.  We sometime split larger ones into segments, ranked each segment, and then combined these for overall ranks for the stand.

Montane grassland in Gillette Canyon was ranked A for native cover, invasive exotics

and native increasers; B for soil surface condition (old roads).  Dead pine on the slopes

burned in the 2000 Jasper Fire.

It is a real challenge to design a protocol in the office that will work in the field.  The range of variation being assessed, and the ability of surveyors to confidently recognize categories within that variation are the main confounding factors.  Prior experience with Black Hills montane grasslands made development easier.  But I was sure we would have to make adjustments in the field.

During the first week of field work, three of us worked together and put the protocol to the test, independently evaluating each grassland we surveyed.  In most cases we came up with the same condition ranks, but not always.  I made a few adjustments to categories to better represent the range of variation seen in the field, and to make the categories easier for us to distinguish with confidence.  We repeated the experiment several weeks later, and tweaked one of the metrics a bit more.

As I went from grassland to grassland filling out forms, I became more satisfied with the condition evaluation protocol.  The results were consistent with my gut feelings expert opinions, and the procedure was quite doable.  Even better, now as I write I feel confident about our results and recommendations.

If You-the-Reader happen to be a vegetation ecologist faced with this type of task, heed these words:

Know thy subject well.

Hold off on data collection and do reconnaissance first to get to know your subject.  Can it be circumscribed easily or are there issues with definition?  What kind of variation will you have to characterize, evaluate?  What are the down-ranking factors?  What can realistically be expected of field workers?  These projects almost always include more area than can be covered in the time allotted, and we race to survey as much as possible.  It is tempting to skip reconnaissance and start collecting data right away, but this rarely works.  Passing ecological judgement is a big responsibility, and it is worth investing the "extra" time.