Tip:
Highlight text to annotate it
X
This is actually a repeat of information
that I gave last week at the Chesapeake
Community Modeling Symposium,
focusing on a project that we have going on
over on the Eastern Shore in the town of Queenstown.
This is a project funded by NOAA (National Oceanic and Atmospheric Administration)
through CICEET (Cooperative Institute for Coastal and Estuarine Environmental Technology). It's a quarter of a million dollar project
where we're using multiple models--
evaluating the use of multiple models--to inform
land planning at the local scale.
And today, what I'll do is I'll start off by providing
some background and context for the project,
and then put that into the framework of the
overall decision-making framework that we used,
at the planning scale.
So I have to tell you here that I'm wearing both hats,
both as a watershed planner, and also a watershed modeler,
so I'm going to move back and forth between those two roles.
And then I'm going to finish up with
really kind of highlighting what I think is significant
from a scientific standpoint and moving forward with
managing the Chesapeake Bay watershed and reducing
non-point source pollution.
So the reason that we're focusing on the local scale
is because almost 90 percent of the watershed
and its shoreline is rural.
And so this is the scale at which local planning
is being implemented, where most of the land
management plans or decisions are being made.
So we really need to understand the connections between
the science and the management at this scale.
I focused on Queenstown, not only because I live there,
but also because it's one of these municipalities
where sewer and water infrastructure have really
limited growth over the past 10 years.
And it's close enough to Baltimore and Washington, D.C.
that it's really primed for growth over the next
30 to 50 years.
In fact, Maryland State and the Chesapeake Bay program
think that this area could grow by as much as 20 to 30 percent
over the next 30 years.
So it's about 20 minutes east of Annapolis,
10 miles east of the Bay Bridge.
Very close to D.C. and Baltimore.
The town is in the process of developing its comprehensive plan,
and one of the main priorities of the community is
to preserve open space around the town,
and preserve the town's rural character.
So what the planning commission came up with was the idea of
accepting growth, consolidating that growth
at a higher density closer to the existing municipality,
and in return, preserving the open space
in the surrounding town planning area.
Now, to accomplish this we need to appease the land owners
and assure them that they're going to get the full value
of their land.
And we also have to convince the citizens
to accept the additional growth in the expansion of high density
housing around the existing municipality.
So, what the planning commission did
was come up with two different growth scenarios.
One is a distributed growth scenario--
let me back up a second and point out
that the top diagram highlights the existing
municipality in red.
The second, middle diagram would represent
full buildout under county zoning.
So you can see the difference in development area between
the county distributed growth and the current
growth scenario.
And then the third, sorry, the second growth scenario
would be the consolidated growth option.
Again, where we have a relatively high density
of development occurring adjacent to the existing
municipality and we manage, or we have say over
what's occurring in the outlying areas,
especially in terms of agricultural practices.
And you can see here that the difference in
development, agriculture, and forest area
compare--we can see that development almost doubles
under the distributed growth scenario
and remains largely unchanged, only a three to five percent increase
in development area under the consolidated growth option.
So, back to wearing my planning hat,
in addition to assessing non-point source pollution,
we also wanted to look at open space
and wetland and forest protection.
But we also needed to consider other factors,
like traffic impacts, water and sewer infrastructure,
school population impacts, job creation, and local economy.
And these are all equally important to protecting
the environment, and preserving
the character of the town.
I just wanted to take the opportunity to also highlight
that under a recent bill passed by the State,
the Maryland House Bill 1141, the towns are required
to prepare a water resource element.
And this, I think, is a really great opportunity
to intersect the watershed science with local planning,
because it requires the towns not only to
assess sewer and water infrastructure impacts,
but to do these non-point source assessments.
We wanted to use multiple models in part because
of other work that we're doing at SERC (Smithsonian Environmental Research Center)
looking at comparing different watershed models.
And this is actually the third project where
we're using multiple models to assess watershed impacts,
and this is a highlight from one of our studies
looking at published models in the Patuxent watershed.
You can see here there's an excerpt looking at
annual average total phosphorus loads,
in three sub-basins of the Patuxent watershed.
We have Laurel, West Branch, and Bowie.
And the different colors represent different models
including the Maryland Department of Planning's
non-point source assessment model,
SERC's empirical and landscape models,
the SPARROW (Surface Water Quality Modeling) set of models from ?87, ?92, and ?97,
the Chesapeake Bay's Phase Four and Five model outputs,
and also the Patuxent Landscape Model.
And what I want you to realize or notice is that if you look at
any one color you'll see that most often
a model does really well in one location,
but relatively poorly at another location.
So, for example, we can look at
the SPARROW ?97 model, you see it does a really
nice job of predicting total phosphorus loads at Laurel,
but a relatively poor job predicting phosphorus loads at Bowie.
So this really convinced us of the value of using multiple models.
It provides a formal framework for accounting for
the structural uncertainty of the models.
And it also provides a stronger basis,
or more coherent basis, for communicating
the uncertainty to the land managers.
Okay, so for our Queenstown application we used
preliminary output from the Phase Five model.
And we used the Maryland Department of Planning's
and Environment's non-point source assessment model,
which is essentially a down scaling algorithm
for applying the Bay Programs' data.
We also developed a SWAT application
(Soil Water Assessment Tool), and also used the Generalized
Watershed Loading Function (GWLF), both developed independently.
And then finally we used the Long Term Hydrologic
Impact Assessment (L-THIA) model, which predicts
annual average loads.
And this is a fairly popular model with the planning community,
which it why we also included it in this analysis.
For the temperately dynamic models,
including SWAT and GWLF, what we did is we calibrated
the models with three basins including Greensboro
and two smaller basins where the Smithsonian
has collected monitoring data.
And then we used Ruthburg, and two additional watersheds
to validate our models.
And then we could transfer the parameters
used from that calibration to look at land cover impacts
in our target, or applications area.
For our application area, what we did is we used
two meter DEMs (Digital Elevation Models) to delineate the watersheds,
and then we intersected these hydrologic units with
the planning area so that we ended up with
hydrologic units that we could use to communicate
between the watershed models and the land planning scenarios.
And this is just an overview of the model that we used
for the calibration and validation.
You can see here: (1) the time period for which
we have observation data ranging from 2 to 20 years.
For the model applications we used 10 meter DEMs,
2000 land cover data from RESAC (Regional Earth Science Applications Center),
and the SSURGO (Soil Survey Geographic) soil database.
So the first set of results that I'm going to share with you
are derived from the Maryland Department of Planning's
non-point source assessment.
And you can see I have total nitrogen on the left,
and annual average total phosphorus loads on the right,
with the current or baseline conditions in the top panel,
the county or distributed growth conditions in the middle panel,
and the consolidated growth impacts
on the bottom panel.
And two points to recognize:
first, relatively high total nitrogen and total phosphorus loads.
And this is the class break out here is based on
what SPARROW's been using to communicate their results.
So these nutrient loads are on the higher end
of the nutrient loads predicted by the SPARROW model.
The second pattern that you probably realize
is that there's not a lot of difference between
the three scenarios when you're looking at either
total nitrogen or total phosphorus.
And this will become even more clear
if we look at the tabular data.
So what we have here: total nitrogen inputs
on the top, outputs, sorry,
to the lower Chester River, and on the bottom
we're looking at outputs to the upper Wye River.
I should've pointed out that Queenstown straddles
both of these watersheds.
And then what we're looking at is the total discharge
of nitrogen predicted for different
buildout scenarios:
current, county zoned buildout, consolidated growth
with no agricultural BMPs (Best Management Practices), and a consolidated growth
with agricultural BMPs.
And if you look for any single model,
you see that there's not a lot of difference
in the total nitrogen loads predicted in either watershed.
And this becomes really clear when you compare predictions
for any single scenario among the different models.
And clearly, the range of variation in the predictions among the models
is much greater than the range of variation imposed by
the different land cover scenarios.
And the same pattern emerges when we look at
total phosphorus loads for the
lower Chester River and upper Wye River.
Again, we get relatively small variation among the different
land cover scenarios in comparison to the variation
predicted from using multiple models.
So when we bring this back into the decision-making framework,
and think about what are these watershed impacts
for the different alternatives that the planning commission is considering,
what ends up happening is because there's relatively
small differences among these different scenarios,
these objectives really fall out of the decision-making framework,
and these other considerations play a much bigger role
in evaluating which alternative to consider.
And what's really hard, or what really kills me,
is when you think about the complexity,
the time, and the resources that went into developing
these predictions in comparison to the tools that are being used
to evaluate the alternatives for these other considerations,
in some cases, it's just a simple ranking
of what we think is going to have the most impact
versus the smaller impact.
So, coming back, putting back my science hat on,
we remain convinced that using multiple models
is a really effective way for understanding
the uncertainty in the model predictions.
And I think what we've done
by using multiple models is we've improved
our understanding of where the uncertainty
in the models lies.
And we get concrete ideas of where to move forward,
where to target research,
how to interpret the model results from using these
combinations of models together.
The second point that I wanted to make is that
I think that our results and our effort shows
that these downscaling tools that the State is using
is really ideal for this level of planning.
Right, so clearly investing in the development
of a complex model for the local scale
is probably not the best use of our resources.
Maybe the way we can move forward
is to think about using multiple models
to come up with a probable range of loading coefficients
for different land covers and best management practices.
Those are the two main points.
The third point that I wanted to put on the table
as my experience is that it's been really
interesting to realize that the citizens
and the planning commission largely think of
agriculture and open space as equal.
And we all know that that's not true,
and something that we should think about how to address.
And so with that, I'll end there.