The War of Models in the Bay

The Big News in the Bay for the week is that USDA released a report yesterday that touted conservation programs already in place in the agriculture industry, many of them voluntary, as having a beneficial effect on the Bay:

"Agriculture plays an important role in protecting water quality and maintaining economic stability in this watershed," said Dave White, Chief of the USDA Natural Resources Conservation Service, as he announced the study results today. "This study confirms that farmers are reducing sediment and nutrient losses from their fields. Our voluntary, incentives-based conservation approach is delivering significant and proven results. This study will help us improve our conservation practices in the Chesapeake Bay area."

The study also shows that there are opportunities for further reductions of sediment and nutrient losses from agriculture by focusing conservation activities on the most vulnerable acres. Well managed farmland is among the best land uses for sustaining natural resources in the watershed. Conserving working lands will be instrumental in meeting objectives for a healthy Chesapeake Bay.

Key findings of the Conservation Effects Assessment Project (CEAP) study include:

• Conservation practices have reduced edge-of-field losses of sediment by 55 percent, nitrogen in surface runoff by 42 percent, nitrogen in subsurface flow by 31 percent and phosphorus by 40 percent.
• Targeting enhances effectiveness and efficiency. Use of additional conservation practices on acres with a high need for additional treatment can reduce per-acre sediment and nutrient losses by more than twice that of treatment of acres with low or moderate conservation needs.
• Comprehensive conservation planning and implementation are essential. The study shows that the most significant conservation concern on cultivated cropland in the watershed is the loss of nitrogen by leaching and overland flow. Suites of conservation practices that include soil erosion and comprehensive nutrient management are required to address soil erosion and nutrient losses simultaneously.

The report is based on a computer model similar in many ways to EPA's Chesapeake Bay Model, which differs somewhat in it's input data, and parametrization.

To paraphrase Tom Lehrer, "Models are like sewers.  What you get out of them is what you put into them."

The USDA model has been well received by farmers and their advocates, as it somewhat mitigates the apparent impact of agriculture on the Bay, and they are using is as an excuse to cast doubt on the EPA model and the resulting TMDLs that put a considerable emphasis on reducing nutrient inputs into the Bay by putting the squeeze on agriculture.

Such attempts were not well received by advocates of EPAs approach, or even more stringent restrictions.  After all, they don't grow Chilean Sea Bass on the Eastern Shore...

I will offer one rebuttal that had to be largely written in anticipation of the USDA report, an article from Chesapeake Bay Foundation, citing one of their own experts, Dr. Beth McGee as to why the USDA study was erroneous:

...The LimnoTech report levied criticisms at the computer model used by the Environmental Protection Agency (EPA) to develop the Chesapeake Bay ‘pollution diet’ or Total Maximum Daily Load (TMDL). Specifically, the report compared the Chesapeake Bay Program (CBP) Partnership’s Watershed Model to one used by the U.S. Department of Agriculture (USDA) in its Conservation Effects Assessment Project (CEAP) report for the Chesapeake Bay Region. LimnoTech alleges that differences between the two modeling efforts draw into question the validity of using the CBP Watershed model to develop the Bay TMDL. This contention is completely without merit. Not only is the CBP Watershed Model a fully valid basis for the TMDL, the CEAP report reaffirms the need for agriculture to do far more to reduce its water quality impacts.

The CBP Watershed Model and the CEAP model were developed for two different purposes. The CBP Watershed Model was created as a management decision-making tool to assist with the development of the TMDL and includes comparable information about multiple pollution sources. The CEAP model is more narrowly focused on evaluating the effects of conservation practices on cropland. Because they were developed independently to achieve different goals, it is not surprising the modeling framework and several model parameters (e.g., hydrology, time frame, spatial scale) differ. Hence, comparing the models is like comparing apples to oranges.

At its core, the LimnoTech report is an attempt by national agribusiness lobbying groups to discredit the Chesapeake Bay TMDL and delay efforts to clean up the region’s rivers, streams, and the Chesapeake Bay. The Bay TMDL is a scientifically-based tool developed over a decade in collaboration with numerous federal, state, and academic partners using a state of the art model that peer reviews have validated time and time again. ANPC’s efforts to undermine the TMDL by attacking the credibility of the CBP Watershed Model distracts us from the real issue that agriculture, like all other sources of pollution, must do more if we are to restore the Chesapeake and the rivers that feed it.

I won't pass judgment on either model; my modeling expertise is minimal, and I have a jaundiced view of modeling in general.  As I have seen over and over, people who go looking for facts in a model tend to find what they expect to (and what they programmed into it).  All models are only approximations of a larger system, much of which evades capture. I'm only going to shout this once: ALL MODELS ARE WRONG, THE ONLY QUESTIONS ARE HOW WRONG AND IN WHAT WAYS!

I would say that the editorial by CBF repeatedly takes a cheap shot, constantly referring to the USDA study as the "LimnoTech report" and implying that they are industrial stooges of the agriculture. I daresay that most of the modeling carried out by EPA is actually done under contract to big environmental firms too.  So it would be fair for the farm organizations to call the CPB Watershed model the "Environ Model" or the "TetraTech model" depending on who did the majority of the work.

The EPA and USDA simply represent two different interest groups within the Federal government.  EPA largely represents the environmental lobby, the regulate everything, regardless of costs approach, while the USDA represents the "don't regulate us, we're too important" lobby for farmers.

I'm willing to bet that in the long run, the facts will show that neither model catches the truth all that well, and that the truth is likely to lie somewhere between the two.