Projects > National Capital Region Network Vital Signs Monitoring

groupings for Integrated Assessment


We have had bits and pieces of this conversation - but I thought we could try a posted discussion to keep it moving along, let's try anyways and see if it works?

For our ROCR data which is coming in (and then all the parks) - one question still remains as to how to divide the park(s) for comparative assessments. The options are 1. geographic or physical division (eg watershed) or 2. biologic/community/habitat division (eg vegetation class).
To follow on from discussions with Bill, Shawn, Todd and Lisa on this  - I wonder if we may get most information by using both approaches.. and the reason is that I think they address different questions and may therefore provide different  information.

To answer the question  "what is the long term pattern in parks ecosystem status over time" requires a fixed, objective grouping (ie watershed or other wise geographic). One of the threats is encroachment, so using a land use classification doesn't help as an area may changes classification from 'woodland' to 'urban' (for eg) - so that while the woodland section is still healthy, it has been reduced in size and turned into imporvious surface. A fixed geographic basis will show a reduction in health status for that watershed as it (for example) increases in impervious surface. These geographic areas may have different metrics going into their ecosystem assessment within a park and certainly they will be different between parks.

But an equally valid question for a resource manager may be "which parks in the system have the most intact woodland habitat". This would be difficult to determine from a purely spatial assessment as any given watershed will have multiple habitats. In this case, it may be more useful to create for each park a table with the habitat types for that park, the area of each type, the percentage of the park under that particular habitat type and the ecosystem status (health index) value for that habitat type within that particular park. After repeated assessments, this would show changes in area over time, but also the relative intactness of that habitat type within that park. Presumably the aim would be to use a standard set of parameters for a given habitat tpye (eg for grasslands: deer abundance, water nutrients, invasives etc, for woodland: community structure, connectivity etc).

Just some thoughts anycase to keep this ball rolling - as the data is accumulating!

I agree with Tim that both types of divisions will be useful in this assessment.   This may make life more difficult for me, having to do twice as many assessments, but hopefully it will be more useful scientifically.

If we do decide to go this route, we need to decide which divisions we will be using for the assessment.  For geographic divisions I think we should use the same subwatersheds we are using for the aquatic data.  Using this division for the terrestrial assessment may allow us to correlate the terrestrial data with the aquatic data for the overall assessment.  I know we have decided to use habitat types for the terrestrial data, but we have not decided which set of habitat types to use.  That should be the next step while we are waiting for the remainder of the data to be posted.

Using a standard set of parameters for each habitat type is also a good idea.  This way we can compare between parks without too much difficulty.  Once we decide which habitat divisions to use, we can then determine which vital signs fit within each habitat.

Nice start Tim and Lisa.  My thought with respect to this topic is to keep it simple, and I present this more from a political tack than a scientific one.  The spatial extent of habitat/geography is the most easily understood, but I also agree that you need measures of condition.  From a management perspective, I think you can answer this via two simple questions: 1) How much do I have,? and 2) What condition is it in?.  One could also add a third practical question: Can I do anything to fix a problem that exists?

Defining spatial extent will address question 1 and risk assessment (the stressors that Tim mentions are good examples) will address number 2.  Number 3 is really most relevant, in that it pertains to adaptive management and, in a sense governs the first two (e.g. a park cannot add area to improve an ecosystem beyond legislated boundaries and a park may not be able to directly reduce nonpoint-source pollution). 

So, addressing "How much do I have?".  Subwatersheds seem to be logical.  I would propose that we use Anderson-Level 2 or higher for the terrestrial side, as this level is pretty easily classified from remotely-sensed imagery and would related directly to Todd's work.  Although I applaud Lisa's suggestion that we relate terrestrial and aquatic units, however, there has been quite of bit of discussion on the topic from the statistical community about how the inferences just don't mesh.  For example, a water monitoring protocol will have a sample design based on stream order and placement within the watershed, but a terrestrially-based protocol (let's use birds, or forest vegetation) will have a completely different sample design, and consequently, statistical inference.  Basically, it is very hard to terrestrial indicators to say something meaningful about aquatic ones, both statistically and (sometimes) logically.  I suppose stream buffers or some "interface indicator" might be an exception.

Addressing "What condition is it in?": I would suggest going to thee literature and finding what's out there in terms of indicators that have been assessed by stream order or broad habitat types.  I am pretty sure there are sources out there that give reach-specific water chemistry parameters (1st-order DO is this, 4th-order is that...).  However, I think we will have a hard time finding terrestrial indicators that relate to very specific habitats (e.g. an oak-hickory forest should have x deer per ha but a floodplain forest should have y deer per ha).  And even if we have data to support these differences, would a manager be able to use that information?  That is why I think picking the broadest, most general, terrestrial habitats might be a good starting point.  Todd can correct me, by I think Anderson 2 would give us only 5 or 6 habitats at ROCR. 

In closing (my cigar is almost out), it might help to think about what we want to say at the end and let that drive our first set of groupings.  My take on it is that we want to say "ROCR has lost x amount of forest since xxxx and the condition is y because of certain thresholds (exotic cover, deer density, bird diversity...).  First order streams are this, second order are that, and so on, based on comparison to EPA or DC regulatory standards.  Lisa - think about exactly what you want to be able to say about ROCR and that will determine what units you choose.  If we don't have the data yet, that's OK.  Keep is relevant, keep it simple - we can always subdivide when we have more data.


[0] Message Index

Go to full version