Too much of a good thing: predicting nitrate
contamination in the Ipswich River
Professor Gil Pontius's research
Environmental analyst Gil Pontius and his students are hot on the heels of the ripple effects
of changing land use. Using
computers and geographic
software, they simulate and make predictions about the environmental impact of land use changes.
- Pontius and undergraduate Stephen Aldrich '02 participated in a land use change modeling contest.
- Aldrich and Pontius examined how
varying levels of detail in source information can affect predictions of
deforestation (see poster)
- Pontius recently supervised the landuse research of six undergrad HERO fellows during summer 2001 and 2002.
- Emily Shusas '02 and Diana Huffaker '03 worked with Pontius to pursue their interests in changing landuse.
Below, read about the work Pontius
and his colleagues, including graduate student
Laura Schneider, have undertaken on the impact of land use change in Ipswich, Massachusetts, an area
north of Boston. Pontius obtained
funding from the National Science Foundation to employ undergraduate Steve Aldrich to work on the project in summer 2001. (Read an online interview with Steve.)
The problem
Nitrates (NO3) are
chemicals that can regulate whether or not there is a healthy balance between plant and animal life in
river systems and wetlands. Nitrates are necessary food for many kinds of
plants, but when levels of nitrates are too high, a condition called
eutrophication results. Aquatic vegetation becomes too abundant and hinders
aquatic animal life.
Nitrates are
present in sewage and fertilizer, and thus soil in residential areas is likely
to have a high level of nitrates. When rain washes soil from these areas into
rivers and streams, the nitrate levels present in the water increase and result
in eutrophication. As more and more forested land gets converted to residential use, it
is likely that the level of nitrates in surrounding water systems will also
increase.
Pontius and his
colleagues suspect that increasing levels of nitrates are likely to be a
danger to the Ipswich River watershed in Massachusetts over the next century,
where there is increasing pressure to convert forested land to residential use. The Ipswich River runs north of Boston, Massachusetts and flows into Plum Island
Sound. The Sound and its watersheds constitute a Long Term Ecological Research Site funded by the National Science
Foundation.
Plum Island itself is a
coastal barrier island known to
New Englanders as an area of great scenic beauty. Visitors go there to enjoy the
lovely beaches and to bird-watch in the Parker River Wildlife Sanctuary.
Goals
Pontius and his
colleagues are trying to accomplish two things with their research:
- They hope to
predict possible future levels of nitrate pollution in the Ipswich River
watershed by looking at how much forested land is likely to be converted to
residential use and where the conversion is most likely to happen
- They hope that
the methods they develop for making this prediction, if successful, can be
used to make similar predictions for other endangered areas
Information sources
- Information
on land use comes from maps provided by MassGIS
for 1971, 1985 and
1991
- Additional
information on wetland location is derived from aerial photographs of the
region
- Students have done
additional research to create maps that indicate areas prevented by law from being converted to residential and
commercial use
How they did it
Pontius and his colleagues have created a computer program
that is used to simulate land
use change between two time periods. The simulation was tried first on the
period from 1985 to 1991. Results were compared with the actual 1991 land use
map to see how accurate was the prediction. Then the simulation was repeated for
the period 1991-2101.
1. First,
a suitability map was created. Each forested location on the map was assigned a
number indicating how suitable it would be for residential land use. The
simulation assumed that areas of higher suitability would get converted to
residential use first.
Suitability was based on
- distance to major roads
- distance to minor roads
- distance to rivers
- distance to existing residential areas
The closer a location was to any of these features, the more suitable it was
considered for residential development.
2.
Every location that had a legal restriction prohibiting conversion to
residential/commercial land use was automatically given a very low suitability
number.
3. The
amount of forested land converted to residential use was computed for the
14-year period from 1971 to 1985. That rate was then used to estimate how
much land would be converted for just a 6-year period, the time between 1985 and
1991, and then for a 100-year period, from 1991-2101.
4. They
ran the simulation on the computer, using information about suitability and the
amount of forested land that can be expected to be converted to residential over
a given period of time.
Conclusions
- The simulation predicted
that approximately 45% of land
forested in 1991 will be deforested by 2101.
- The simulation
predicted that nitrate
release has the potential to double by 2101.
- The simulation did an
excellent job of predicting over a 6 year period (1985-1991) the amount of
deforestation that would take place, and a fair job of predicting
where the deforestation would take place. Time will tell whether these methods
will do as well over a period of 100 years
Summary poster
Click here
to see the poster that Pontius and his colleagues created to present their
research.