AIRMoN Dry Deposition

Background.

ARL is a leader in the development and operation of dry deposition networks. Since 1984, the Atmospheric Turbulence and Diffusion Division in Oak Ridge has been operating a network specifically designed to get around the major problem confronting dry deposition monitoring activities -- there is no existing method suitable for routine direct measurement. The nested network that was developed consisted of a small number of research sites supporting a larger array of stations making simpler but more routine observations. The Dry Deposition Inferential Method (DDIM) that was developed remains the central routine analytical tool of the ongoing NOAA dry deposition trial network, now identified as the dry deposition component of the Atmospheric Integrated Research Monitoring Network (AIRMoN). This network started with six sites; thirteen stations are now operating (as shown in the AIRMoN network map elsewhere in this series of documents).

Why NOAA, why now, and why ARL?

Air-surface exchange is a specialty of ARL. Monitoring and coupling with research in an integrated fashion is also an ARL specialty. These two interests and specializations combine in the case of dry deposition, where it is the desire to derive estimates that are scientifically credible even though there is no universally-accepted methodology to make the measurements. In practice, dry deposition of some compounds is about the same as wet, but equality of the two would be unexpected and fortuitous. We know, for example, that the small particles that carry atmospheric radioactivity are deposited about 90% by rainfall, leaving only about 10% for the dry deposition contribution. For many trace gases, the dry contribution is now recognized to be much higher.

The question of dry deposition contributions to ecosystem health requires urgent attention, since the matter is high on the horizon of policy makers and regulators. At present, the relevant science is beginning to become understood. It is the intent of the NOAA programs (conducted by ARL, and also including programs conducted by ARL for the EPA) to provide the best possible quantifications of the rates of dry deposition at selected sites. It is not the intent to provide such data at all locations that are specified on other grounds, since the techniques now available are tentative and exploratory, and are not yet widely applicable.

Network Investigations of Dry Deposition

The NOAA/ARL Atmospheric Integrated Research Monitoring Network (AIRMoN) is in two halves -- one wet and the other dry. Both halves are made up of nested arrays of simple and sophisticated stations. The intent is to conduct research as necessary to improve the measurements and the scientific understanding of controlling processes at the more advanced sites, and to conduct field tests of the methodologies that are developed at the larger array of simpler locations. In the case of AIRMoN-dry, the program is set up with a small number of "CORE" sites, which service a larger array of "satellite" stations. AIRMoN-dry has been in operation since 1985. Dry deposition rates are derived by applying an "inferential model" to measurements of air concentrations of the chemical species of interest, using as input measurements of key atmospheric and surface variables known to be indicative of the controlling processes. The dry deposition estimates that are yielded are thus the products of a site-specific model, driven by local data. It is basic to this that the estimates derived are themselves site-specific. This is not a program designed to yield areal averages, or even to provide estimates of areal averages through time smoothing. It is now clear that wet deposition data obtained at a single location are indeed representative of the surrounding region if the site is adequately selected. This cannot be the case for the dry deposition contribution, since the controlling factors are surface-driven and are not themselves regionally representative. This concept is conveniently summarized in a single statement -- wet deposition is an ergodic process, whereas dry deposition is not.

Coupled with the AIRMoN operation is the EPA-sponsored Clean Air Status and Trends Network (CASTNET), which is the present-day continuation of the EPA National Dry Deposition Network initiated under the National Acid Precipitation Assessment Program in 1988. The same inferential methodology mentioned above is the underpinning of the CASTNET dry deposition program. The EPA and the NOAA activities are wholly collaborative: both operations are structured to apply new understanding as rapidly as possible, and to quantify the uncertainty associated with the indirect (inferential) estimates that the routine monitoring networks produce. In simple concept, the Collocated Operational Research Establishments ("CORE" stations) of AIRMoN-Dry provide a continuing infrastructure for exploratory research. The much larger CASTNET array of the EPA is the routine network in which the results are applied. The tier of simpler "satellite" stations of the NOAA AIRMoN-dry array serves as a transition, testing the transferability from the research environment to routine application.

The Dry Deposition Inferential Method.

Interest, both national and international, continues in the NOAA Inferential Method, initially developed under NAPAP auspices for estimation of dry deposition fluxes. Requests have been received for details of the technique and for the latest version of the inferential model for dry deposition velocity from organizations that are now too numerous to count, from universities and other government agencies to foreign organizations such as in Central Europe, Spain, and South Africa.

In the last two years, a new multi-layered model has been adapted for dry deposition inferential application, replacing the initial "big leaf model" that had its origins in agricultural meteorology. The new model provides a greatly improved capability to simulate the effects of several natural variables known to be important in the dry deposition context. However, some tests comparing the two models indicated little improvement while the newer model was in its early stages of development (for example, results obtained in Europe as part of the EUROTRAC program). Field tests to study the performance of the multi-layered models have been informative. Recent results indicate that a key requirement is to assimilate data as required to get the surface water budget correct.

Research Activities.

The major goal of dry deposition research conducted by ARL scientists relates to the need to identify and understand the processes that cause dry deposition, in order to quantify dry deposition rates at locations where direct measurement is not possible. The focus is on the improvement of models, whether for site-specific application using local observations of key variables as input or for regional application using model "data" fields to drive the deposition routines. ARL presently focuses its attention on

• the development of systems for quantifying dry deposition,
• the measurement of dry deposition using micrometeorological methods,
• the development of techniques for assessing air-surface exchange in areas (such as specific watersheds) where intensive studies are not feasible, and
• the extension of local measurements and understanding to describe areal average exchange in numerical models.

Improved estimates of dry deposition rates for sites in the NOAA AIRMON-Dry network were generated.

Testing of DDIM Predictions

There has been a long history of field tests of the methodsn used to infer dry deposition rates from routinely collected data. Most of these tests have been conducted at sites of the AIMoN-dry array -- the so-called CORE sites (State College., PA; Oak Ridge, TN; Argonne, IL; and recently Bondville, IL). It was on the basis of the experience gained in these studies that the original "big leaf" single-layer model was recently replaced by a multi-layer version. The reason for the change was rthat the ranges of conditions encoutered in field studies demonstrated that a single version of the big-leaf model could not be expected to work well everywhere (although clearly it could be "tuned" to local conditions if adequate local data were available).

It has been a well-recognized philosophy of the team developing these models to make their products available immediately, to all who ask for them, on the basis that these techniques are still in development and that progress towards a high-quality product will be hastened if many groups give attention to the need, independently. In practice, ozone is frequently seen as a key indicative variable. Early studies conducted in Europe with the single-layer model indicated that the model worked well for about five days out of ten; for the others it was way off. More recent studies have indicated that for a pine forest in Holland the single-layer version worked better than the multiple-layer. Clearly, there is more work to do.

The ARL team at Research Triangle Park, working with colleagues at Oak Ridge, has developed a movable system for direct measurement of dry deposition fluxes. A first test of this system was conducted near Beaufort, NC. The system provides direct eddy correlation measurements of sulfur dioxide, ozone, and carbon dioxide fluxes, and gradient measurement of nitric acid flux. The system also measures the surface energy budget. The system was subsequently deployed at the Bondville CASTNET (and also AIRMoN and ISIS) site in Illinois, and later at a coastal site in New Jersey. Deployment at other sites is planned. In essence, the flux data obtained have been used to assess uncertainty and to improve the inferential dry deposition models being widely used in analysis and modeling of dry deposition.

The ASMD team has also conducted an evaluation of existing dry deposition algorithms for gaseous pollutants to identify an algorithm for implementation into the ISC-COMPDEP (Industrial Source Complex - COMPlex terrain DEPosition) model. Model predictions were compared against O₃, SO₂, and HNO₃ field data sets. Sensitivity tests showed that the models were most sensitive to land-use type and time of day (day/night), so the data sets were stratified based on these classifications for use in the evaluation. The question of how to handle the fact that dry deposition processes differ greatly from day to night has yet to be well resolved.

Collaborations

Three ARL teams are heavily involved in these studies -- Oak Ridge, Research Triangle Park, and Silver Spring. In addition, a large number of other organizations are partners with ARL in the research that underpins AIRMoN-dry and its developmental programs:

University of Vermont	Vermont Monitoring Cooperative
Pennsylvania State University	US Geological Survey
The University of Maryland	Environmental Protection Agency
The University of Illinois	Department of Energy
The State University of New York, Albany	Argonne National Laboratory
The State University of New York, Syracuse	State of Vermont
University of Maine	National Park Service
US Army