Sunday, July 05, 2009

Abuse of science and logic by the National Corn Growers Association

The National Corn Growers Association released a report arguing that there is no connection between the use of nitrogen fertilizers on corn in the Midwestern US and the seasonal “Dead Zone” in the Gulf of Mexico.

There is no point mincing words about what this “analytical white paper”. It is the corn equivalent of irrational climate change skepticism. This is one truly shoddy piece of work. I encourage others in the scientific community to respond either independently or to append the critique offered here.

First, let’s review the actual science.

The “dead zone” in question, discussed many times before on this blog, is generated most summers on the continental shelf of the northern Gulf of Mexico. Nutrients originating in the Mississippi River Basin in the spring fuel the production of algae (primary production) in the surface waters along the continental shelf. The algae die and sink to the bottom, or something else eats the algae and the fecal matter from the something else sinks to the bottom. All that organic matter needs to decompose, and the process of decomposition (respiration) consumes oxygen. So the bottom waters on the continental shelf during the summer become very depleted in oxygen, or “hypoxic”.

Scientific research over the last few decades has shown that the increase in nitrogen flow from the Mississippi and neighbouring Atchafalaya Rivers since the 1950s has driven the development of these large seasonal periods of hypoxia. The evidence comes from basic ecological theory on nutrient limitation, lab experiments, tracking of the Mississippi River plume, long-term data analysis, sediment cores, isotopic analysis and mathematical modeling. While other nutrients like phosphorus and silica are important, nitrogen is the primary culprit.

There are many possible explanations for the increased flow of nitrogen out of the Mississippi-Atchafalaya River Basin (MARB) including fertilizer use, manure use, NOx emissions from cars and sewage. A simple nutrient budget shows nitrogen fertilizer use in the MARB has increased 20-fold since the 1950s. And today, most of that nitrogen fertilizer is applied to corn fields. Measurements and mathematical modeling of nitrogen loss from corn fields show that corn production is a primary source of nitrogen to the Mississippi and Atchafalaya Rivers, and hence, a primary driver of the development of what's come to be called the “Dead Zone”.

The author of the NCGA report (from the consulting firm StrathKirn Inc.) attempt to counter the mass of scientific evidence with the following largely baseless and unscientific arguments. Basically, he throws a bunch of stuff at the wall to see if anything sticks. I’ll go one by one through the report's chain of five incorrect and comically inconsistent assertions:

Assertion #1: Oxygen levels on the continental shelf are not low in comparison to other parts of the ocean.

This is misleading and irrelevant. First, the large regions of upwelling in the open ocean have low oxygen concentration due to high primary production. There’s no sense in contrasting the naturally and persistently low oxygen levels in the eastern Pacific to the intermittent, seasonal hypoxia on the continental shelf of the Gulf of Mexico. Second, even if there were some sense in this comparison, the data resolution of these maps is far too poor to capture a hypoxia area, which, while among the largest in the world, is still at its largest on the order of 20,000 km2 [here’s a test – can you clearly delineate New Jersey on that map?]. The global map of marine nitrogen concentrations is even more ridiculous. The data is far too coarse to capture the plume of the Mississippi River.

Assertion #2: Hypoxia doesn’t affect the fishery (not there is any hypoxia).

The report shows no change in fish catch over the years. As Steve Carpenter of the University of Wisconsin mentioned in an e-mail, the problem is the report analyses data on fish landings, not fishing effort. The boats may come back with the same weight in fish – but it takes more time and money to get those fish.

Assertion #3: Nitrogen from the Mississippi and Atchafalaya doesn’t cause the hypoxia (not that the hypoxia affects the fishery, or that there is any hypoxia in the first place).

This argument is advanced through a series of graphs relating annual nitrogen export, annual river flow and the annual extent of the hypoxic zone. There are a number of problems here. The nitrogen and flow data are shown only since 1985, despite data existing back to the 1950s. If the graph went back thirty years, you’d see the 2-3fold increase in nitrogen export occurred between the 1950s and the 1980s. Instead, the author reports no evidence of a trend in nitrogen of hypoxia since 1993. That’s not the issue – the issue is the hypoxic zone began growing large in the 1980s because fertilizer use increased between the 1950s and the 1980s, and further increases in corn planting, say for ethanol production, may further increase the average annual extent of hypoxia.

The other glaring problem with this argument is that the report uses no statistics whatsoever. For example, after a chart of nitrogen export and hypoxia extent since 1985 is this unsupported passage:

Again, there appears to be an association between water flow and the amount of nitrite (NO2) plus nitrate (NO3), but these do not relate well to the size of the hypoxic zone (except that they are all low in the year 2000). Thus, many of the statements about the relationship between water flow, nitrogen, and the size of the hypoxic zone are inaccurate.

Some actual statistical analysis, or frankly, just eyeballing the graph, would suggest that there is a significant relationship between the annual nitrogen export from the MARB and the annual extent of the hypoxic zone. It is not a perfect one-to-one relationship between nitrogen and the extent of hypoxia because of how the weather effects mixing of oxygen in the Gulf, the load of other nutrients and a myriad of other mitigating factors. If the author had done any research, they’d find proper statistical analysis and explanations in dozens of published papers, including this one of from my own work, a 2007 paper in Limnology and Oceanography:

Between 1985 and 2004, there is a significant relationship (r2 > 0.61) between midsummer hypoxia area and the May + June nitrate flux (Fig. 1). The strength of this relationship is limited by a number of other variables, including the advection of sub-pycnoclinal waters on the continental shelf, summer tropical storms that increase vertical mixing, recycling of N sequestered in shelf sediments during previous years, and the input of other nutrients such as phosphorus (Rabalais et al. 2002; Scavia et al. 2003; Wawrik et al. 2004).

Assertion #4: Not very much nitrogen is applied to corn (not that nitrogen causes hypoxia, or that hypoxia affects the fishery, or that there is any hypoxia in the first place).

The report displays a graph illustrating that non-crop uses of nitrogen fertilizer, like fertilizer used on lawns, is equal to or greater than the use of nitrogen fertilizer on corn. The problem, or I should say, the most glaring problem? It is national data. Over 90% of the corn grown in the US, and over 90% of the nitrogen fertilizer applied to corn in the US, is grown in the MARB. A 1999 EPA report estimated that non-agricultural fertilizer use is only 5% of total U.S fertilizer use - and that percentage of total fertilizer use in the major producing states of the Corn Belt.

Assertion #5: No nitrogen runs off of corn fields (not that much nitrogen is applied to corn, or that nitrogen causes hypoxia, or that hypoxia affects the fishery, or that there is any hypoxia in the first place).

The report proudly claims that the same amount of nitrogen is now removed during the corn harvest (i.e. in the grain) than is applied as fertilizer, so there can’t be any extra nitrogen left over to run off into the river. Fertilizer use efficiency has indeed increased over the years thanks to genetic technology and improved management. In other word, farmers are getting higher yields with the same amount of nitrogen fertilizer. That is positive news.

But the calculation in the paper is full of flaws. To name just one: the contention that fertilizer inputs = crop outputs = no nitrogen runoff only makes sense if fertilizer were the one and only source of nitrogen to the crops. For one, there is the mineralization of nitrogen in the soil – plant matter on the ground is naturally broken down by microbes, a process that released nitrogen from the plant matter to replenish the soil. This is a fundamental part of soil chemistry. The whole reason the Midwest is good land for growing corn is the high natural mineralization rates!

Final take-home message of the report: The US has a lot of golfers.

The report concludes that all other analyses are ignoring all the fertilizer applied to lawns and present maps and data to support this conclusion. The calculations are extremely suspect. First, the author assumes that the fraction of land devoted to lawns is greater in the MARB than in the rest of the country. Analysing the lawn data, eyeballing the national map, or simply reflecting about the fact that 4/5ths of the US population live outside the MARB, shows that this is a ridiculous assumption. Second, the report assumes that all the fertilizer not applied to corn, wheat, soybeans or cotton – which amounts to about 25% of annual fertilizer sales - is applied to lawns. This ignores all other crops grown in the United States, as well as all the fertilizer applied to rangelands and forests.

The report goes on to argue:

Since most lawns are cut and mulched there is relatively little removal of N, unlike the grain in corn. Consequently, a major portion of the N applied to lawns may be available for leaching… the net N available for leaching per acre is almost infinitely higher for lawns than from corn.

Not only does this argument incorrectly imply that no plant residue whatsoever is ever left behind after harvest to replenish the soil, it ignores the fact that unlike lawns, many corn fields are artificially drained by pipes or drainage tiles, such that excess nitrogen easily leaches to the nearest stream.

All told, the NCGA report is an embarrassment.

There are some legitimate outstanding questions about the nitrogen-hypoxia problem and definitely some legitimate critiques of the media coverage. In particular, the coverage often gives the mistaken impression that corn is the only source of nitrogen, that the hypoxic zone covers a large fraction of the Gulf of Mexico, that water at all depths is hypoxia, and that hypoxia is a permanent phenomena, rather than a seasonal occurrence. The NCGA could have issued on a report on those real concerns. Instead, it issued this dishonest mess of half-truths and pseudo-science.

No comments: