United States Environmental Protection Agency	Office of Research and Development

National Exposure Research Laboratory
Research Abstract

Government Performance Results Act (GPRA) Goal 4
Annual Performance Measure 265

Significant Research Findings:

Demonstration of a Coupled Plant-Soil-Deposition Model to
Improve Deposition Modeling

Scientific	The EPA operates the Clean Air Status and Trends Network (CASTNET) to

Problem and	monitor the status and trends of air pollutant emissions, ambient air quality, and

Policy Issues	pollutant deposition at over 200 sites across the country. Deposition estimates are

important in assessing the overall pollutant loadings to ecosystems, an important
aspect of ecological assessments. To obtain estimates of deposition flux, measured
concentrations are paired with model-predicted deposition velocities. The model
currently used to predict the deposition velocity is the NOAA Multilayer Model
(MLM). Earlier testing of this model showed the sensitivity of the modeled
deposition velocity to leaf area index. Leaf area index (LAI) and canopy height
are not routinely measured at CASTNET sites and are currently modeled using a
step-function that is based on measurements made in 1991-1992, and 1997 and
depend only on the day of the year. Thus, the same annual leaf-out profile is used
each year, whereas actual LAI values respond to variations in rainfall, temperature,
radiation, etc. Given the sensitivity of the deposition model to LAI, obtaining
better estimates of LAI should provide more realistic estimates of deposition flux.

Research	Site-specific, interannual variations in LAI could be obtained from various sources

Approach	for input to dry deposition models. Remote sensing data offers promise, but is not

currently available at the scales needed for CASTNET. Alternatively, plant
growth models can be used to predict the response of plants to interannual
variability in meteorological conditions. There are many different approaches to
plant growth modeling ranging from simple parameterizations to more
complicated photosynthetically- based models. We selected the Erosion
Productivity-Impact Calculator (EPIC) model for use in this project because of its
relative simplicity and long history of use in the agricultural community. EPIC
predicts plant growth as a simple function of accumulated heat units which depend
on temperature. The deposition model selected for this project is the Multilayer
Biochemical Model (MLBC). The MLBC model is under development as a
replacement for MLM, the current model use for determining deposition velocity
for CASTNET. The plant growth algorithm was extracted from EPIC and
embedded within MLBC. This version of MLBC is denoted as MLBC-PG. In
MLBC-PG, the meteorology used as input for the deposition model is also used
for the plant growth algorithm. Water and temperature stress factors calculated by
the deposition model for determining canopy resistance were used in the plant
growth algorithm. The LAI and canopy height determined from the plant growth


-------
algorithm are used in the deposition model for determining pollutant deposition
velocities.

Results and	The plant growth algorithms from EPIC were imbedded within MLBC-PG to

Impact	provide an optional method for determining LAI and canopy height. MLBC-PG

was run for the CASTNET site in Bondville, IL which was also the site of an
intensive field study during which LAI, canopy height, and deposition fluxes were
measured. The predominant plant type at this site is corn. The values of the LAI
and canopy height calculated by MLBC-PG are in good agreement with those
calculated by EPIC run as a stand-alone model showing that the plant growth
algorithms were correctly imbedded within the deposition model. MLBC-PG was
run for 3 years of meteorological data from the Bondville CASTNET site. LAI
values calculated by MLBC-PG show interannual variations in LAI whereas the
step-function provides the same LAI profile each year. The canopy height curves
are particularly different in that the CASTNET step function decreases the canopy
height during senescence whereas MLBC-PG holds the canopy height at the
maximum value until harvest which is more realistic. Hourly deposition velocities
calculated using the plant growth algorithms were 5-10% different than those
calculated using the traditional CASTNET LAI values, with greater differences
occurring for years with more extreme meteorology. Initial comparisons of LAI
values generated by the plant growth algorithm against the field study data show
substantial differences in the plant growth curves. Adjustments of the input
parameters such as planting data and heat units needed for maturity allow a better
match between the modeled and measured values, illustrating the sensitivity of the
plant growth model to these plant parameters. Additional model runs for other
CASTNET sites show similar results.

These results are a first step in trying to provide better characterizations of plant
growth to deposition models, which can result in better estimates of deposition
flux. MLBC-PG shows promise in being able to predict appropriate LAI and
canopy height curves that are responsive to the meteorology. However, more
work needs to be done to study the sensitivities of this model and to determine the
appropriate plant parameters for use in CASTNET on an operational basis.
Additionally, this project sets the stage for further work in coupling models from
different disciplines.

Dr. Ellen Cooter (NERL/AMD) collaborated on this research, providing valuable
insight in plant growth modeling.

Research
Collaboration and
Research
Products

Future Research The MLBC model is currently under development for use in CASTNET. While

the model has been applied and evaluated for intensive field studies, modifications
are needed to use the CASTNET input data and to specify plant parameters for
network sites. Research is being performed by NERL's Landscape
Characterization Branch on obtaining LAI estimates via remote sensing. The
development of alternative methods for determining LAI and canopy height for
input to the model will depend on client needs.


-------
Contacts for	Questions and inquiries can be directed to:

Additional	Donna B. Schwede

Information	jj.S. EPA, Office of Research and Development

National Exposure Research Laboratory
Mail Drop: E243-04
RTP, NC 27711
Phone: 919/541-2715
E-mail: schwede.donna@epa.gov


-------