I am responding to your NRES Department Report Series 2000-03 (2000-03) and to Greg McIsaac's cover memo of September 12. You prepared these documents in response to the Illinois State Water Survey Contract Report CR 2000-08 (2000-08) entitled "A contribution to characterization of Illinois reference/background conditions for setting nitrogen criteria for surface waters in Illinois."
Your report does provide a few useful comments and suggestions that we have used to prepare, in combination with additional data, an update to our report. However, I find your report to contain misinterpretations and mischaracterizations.
First, I must state emphatically that I disagree totally with your assertion that ISWS CR 2000-08 "has little or no scientific merit." I reiterate later in this response the scientific merit of the report. Other scientists have praised the report.
You state on page 2 that "If K&W's methods are sound, their analysis and conclusions would provide the basis for new scientific research initiatives and would likely require considerable changes in widely accepted scientific theories related to the N-cycle and water quality." We agree with this.
I see our work as a necessary and long-overdue first step in developing an improved understanding of the evolution of nitrogen chemistry in Illinois' soils, waters, and vegetation in relation to landscape changes, natural processes, and human activities. Our report clearly challenges some well-established concepts and beliefs.
I started to work on historical nutrient conditions in Illinois when it became apparent that the National Hypoxia Assessment was not going to address in sufficient detail some of the topics necessary to achieve a comprehensive understanding of nutrient dynamics and hypoxia.
My interest in historical water-quality conditions in Illinois was enhanced when it became apparent that the USEPA had also initiated a process to set new nutrient criteria and standards for the nation. A cornerstone of the criteria development is to establish reference/background conditions. Hence, there is a great deal of commonality between understanding historical nutrient changes as they relate to hypoxia in the Gulf of Mexico and to changes in water quality in Illinois. Interestingly, data used in the National Hypoxia Assessment show that oxygen stress along the Gulf coast increased from the 1700s to 1900s, but that several surrogate indicators of oxygen stress show no increase in oxygen stress since 1950, even though the Assessment concludes that "hypoxia has intensified only since the 1950s."
In our report, we select one period for reference/background conditions, but also state that other periods could be selected. In fact, the USEPA may recommend use of data only over the last 20-30 years to establish reference/background conditions. As far as we were aware, nobody else was seeking to provide a historical perspective on landscape and water-quality changes in Illinois, so it was entirely appropriate to dedicate resources to begin to characterize and interpret these changes.
With a document that draws on more than 500 scientific references - and now about 550 - we have made, as the title of our report indicates, A CONTRIBUTION to the characterization of Illinois reference/background conditions. As we state clearly in the report and in the Executive Summary "This report is the FIRST STEP OF MANY needed to establish the objective, scientifically sound basis necessary for characterization of reference/background conditions in Illinois and for understanding the role of agriculture and other activities on the nitrogen cycle." It was not a goal of the report, as you assert erroneously, to "provide a scientifically sound basis for characterizing water quality background conditions" - it was, as we clearly state, a contribution to such a process. Putting a collection of more than 500 references on the table is clearly much less selective than the National Hypoxia Assessment, which fails to include many of the references and data bases that we present and which are relevant to understanding nutrient dynamics in the Mississippi River Basin.
The main insight that we obtained from our study is the incredible complexity of the nitrogen cycle and recognition of the many natural and human factors that contribute to changes in the nitrogen cycle. The literature clearly identifies up to order-of-magnitude uncertainties in estimating many inputs and process rates.
Our second main insight is the apparent virtual disappearance in modern global and regional nitrogen cycling studies of a hundred years of scientific study representing thousands of experiments on the effects of agriculture in reducing the nutrient content of soils. When we realized this, we had to go back to basics and ask how the conceptual model (paradigm) for studying the nitrogen cycle had changed. This was the origin of what we called the "standing-nitrogen-cycle paradigm" and the "previous nitrogen-cycle paradigm."
Recognizing the complexity of the nitrogen cycle, we are at the Water Survey, as several of you know, preparing an interactive nitrogen web site for release in the next few weeks. We hope that it will be an informational and consensus building tool. Dialogue with and input from other scientists will be encouraged. I am the Principal Investigator for this project.
I will now respond to the major concerns you have raised.
NITROGEN CYCLE PARADIGM
You conclude that "the older literature is consistent with the more recent literature in estimating the magnitude of the anthropogenic additions of N to the biosphere in relation to N fixation...." We agree with this. We also note that you do not challenge our conclusion that recent nitrogen-cycle literature does not incorporate the major anthropogenic decrease in terrestrial-biosphere nitrogen that occurred largely before 1950.
In an article entitled "Soil collection includes a century of history" that appeared in The News-Gazette on September 3, 2000, Ted Peck, archivist of the University of Illinois' Morrow Plots soil samples, is quoted as saying "The settlers farmed without fertilizer. They depleted the soil and they moved on." "They were aware of the depletion they left behind, and they encouraged scientists to study soil to see what they needed to do." In the same article, UI soil fertility specialist Michelle Wander is quoted as saying "We have some wonderful early work in Illinois, with research strategy tied to fertility and history, some very good science." We agree that there are some excellent data bases and some excellent science at the University of Illinois.
HISTORICAL DESCRIPTION OF THE ILLINOIS RIVER AND ITS WATERSHED AND THE ROLE OF WETLANDS.
The Wetlands Initiative (Wetland Matters, 1999) recognizes that there were some 8.21 million acres of wetlands in Illinois in 1780 and that by 1980 this figure had decreased to about 1.26 million acres. Illinois Department of Natural Resources (2000) reports that there are only an estimated 870,000 acres of natural wetlands remaining within the state. In the 1890s there were still large expanses of wetlands along the Illinois River and throughout the basin.
Wetlands, in all their forms, assimilate, store, fix, and denitrify nitrogen and are, by definition, highly eutrophic. The Wetlands Initiative reports that restoring only 407,000 acres of wetlands in Illinois (about 5% of the 1780 wetlands), primarily on flood prone bottomland throughout the watershed, would remove 101,000 tons (80%) of today's nitrate load from the Illinois River.
The Wetlands Initiative cites the loss of the critical nitrogen-removing capacity of wetlands as an important cause of the historical increase in nitrate concentration in the Illinois River from <1.5 mg N/l at the end of the 19th century to average concentrations >5.0 mg N/l in recent years. From this line of reasoning, it is implicit and correct that the concentration of nitrogen in the rivers does not necessarily represent the amount of nitrogen that is input to the rivers and biologically available, but rather represents the amount of nitrogen left over by the plant and animal production process, denitrification, and burial. It is important to look at the amount of dissolved and particulate nitrogen in the water and the amount of nitrogen cycled by the organisms. It can also be concluded that 8.2 million acres of wetlands must have been assimilating, fixing, storing, and denitrifying large amounts of nitrogen per year prior to 1900. Kofoid (1903) reported that in the 1890s Thompson's Lake was supplied with water mainly from the Illinois River, but had an average concentration of total nitrogen about 1.5 mg N/l less than the Illinois River. Could this be due to assimilation, denitrification, and burial of nitrogen in this backwater lake? There is a widespread desire to restore the Illinois River watershed, including reconstruction of large expanses of wetlands. Given that 8.21 million acres of lush, nitrogen-rich aquatic and wetland vegetation were storing and removing large amounts of nitrogen in the 1700s and 1800s, where did such large amounts of nitrogen come from to support the large nitrogen demand of this 8.21 million acres "crop" of N-rich vegetation?
An implication of the loss of the nitrogen-removal capacity of wetlands is the need when studying historical trends in nitrogen concentrations and conducting nitrogen mass balance studies to account for the nitrogen transformation and removal by wetlands (and other environmental factors), when wetlands were prevalent. The amount of nitrogen leaking out of the geosphere to the hydrosphere can not be determined only by the residual nitrogen concentration in surface waters.
Assuming that all 2 million acres of wetland drained in the entire MRB in 1900 were drained exclusively in Illinois, this would leave 6.2 million acres of wetland to take up and transform soluble N from the drainage of the remaining 28.8 million acres of land in Illinois. Given that the concentration of TN in the lower Illinois River was about 24 percent less in the 1890s than the1990s, and assuming this 24 percent difference holds for flux of TN in Illinois' surface waters -- now estimated at about 0.5 billion lbs/yr (David and Gentry, 2000a, b) -- then these wetlands would have to be transforming only 4.1 lb N/acre/yr (4.6 Kg N/ha/yr) from the non-wetland areas to account for the difference between amounts of N in solution in the 1890s versus the 1990s. Put another way, the 6.2 million acres of wetlands would needed to have transformed only 19.1 lb N per acre (21.4 kg N/ha/yr) of wetland to produce a lush 6.2 million acre crop of N-rich aquatic and wetland vegetation every year to account for the difference in total N in solution between 1890s and 1990s surface water.
The most consistent story on the Illinois River itself, as told by eminent people such as Bellrose and Talkington, is that the river, its floodplain, and backwaters had extensive stretches of aquatic vegetation prior to 1900, and that by about 1950 these had disappeared. Mills, Starrett and Bellrose (1966) report that "the first big change in aquatic vegetation came shortly after Kofoid had completed his study, with the diversion of Lake Michigan waters into the river." In selecting Kofoid's 1890s photograph of Quiver Lake to illustrate the aquatic vegetation and organic enrichment, we followed the exact same approach as Bellrose and other scientists at the Natural History Survey (Bellrose et al.,1983, p.9, fig 3) in reproducing a photograph taken during low-water levels to illustrate "luxuriant beds of submerged plants." In fact, we took Bellrose's lead and reproduced the very same photograph of Quiver Lake that Bellrose himself used! Talkington (1993, p. 25) also includes a photograph of dense water lilies. Again following the lead of Bellrose et al., we did not reproduce Kofoid's plate 16 illustrating the lake with considerable open water in another year when high water levels obscured the submerged vegetation. Nor did we chose to use pictures of emergent aquatic vegetation which are more "jungle-like" in appearance. Have you been critical of Bellrose and Talkington in not presenting photographs of open water? Clearly, neither other environmental scientists nor Krug and I see much value in reproducing photographs of open water to illustrate large expanses of submerged wetlands and aquatic vegetation.
You state that "in August, 1821 Schoolcraft noted dense vegetation and very slow moving water in the Lower Illinois River, but he described much of the Middle and Upper Illinois River and its tributaries as being clear." This is true. It appears from Schoolcraft's report that the Illinois River was characterized by large expanses of exuberant aquatic life, unpalatable waters, and clear waters. We state on page 71 that "The contemporary literature takes on face value the words of some that the Illinois River prior to the twentieth century was a clear-water river with nominal erosion..." However, we go on to note that "we believe that what people during the 19th century Illinois called clear water was subjectively different from what people living in the 20th century would call clear water." We provide secchi depth data to illustrate this point. Also, describing water as clear says little about the nitrogen content of the waters.
The conditions of exuberant aquatic life, unpalatable waters, and large expanses of nitrogen-rich wetlands in 1820 can not be attributed to pollution, as the total population of the state at that time was only 55,000 - almost exclusively to the south of the Illinois River (Nelson, 1996). The Wetlands Initiative also makes a number of references to early explorers noting the unrivaled fertility of the soil, the luxuriant vegetation, and the large numbers of catfish, sturgeons and other fish. We note that USEPA (2000a) lists rough fish as characteristic of hypertrophic lakes.
You state that we describe the Spoon River circa 1900 as being "the most pristine of all the Illinois tributaries at the time...." We state clearly on page 72 that this is a reference to Kofoid's 1903 and Palmer's 1903 descriptions and data. To be more specific, below are some quotes from Palmer and Kofoid:
Palmer: "..as there are no very considerable towns situated upon the stream or its watershed, it receives but little sewage and may be regarded as one of the least polluted of the natural streams of the state."
Kofoid: "The small amount [residue in solution, i.e., "the available supply of mineral salts for the phytoplankton as well as some organic materials"] in Spoon River may be attributed to the fact that it is largely uncontaminated surface water of recent origin."
Kofoid: "The sewage systems discharging into this stream are few and but slightly developed, and its chlorine is correspondingly low." [Palmer reported that 'The presence of chlorine in water in amounts exceeding the normal quantity generally indicates that the water has been polluted by animal matters...'] For example, Palmer reported the average concentration of chlorine at Averyville, north of Peoria, to be 30.2 mg/l in 1897-1899. At Grafton, Palmer reported the average concentration of chlorine in the Illinois River in 1899-1902 to be 12.6 mg/l. For the same years, Palmer reported the average concentration of chlorine in the Mississippi River at Grafton to be 2.92 mg/l. [With an average concentration of total nitrogen of 1.59 mg N/l, the Mississippi River could also be classified as eutrophic (USEPA, 2000b), even though it had a low chlorine concentration.] We state on page 72 that Kofoid reported that "the freedom from sewage is evidenced by the low chlorine...". Kofoid reported a mean chlorine concentration of 3.8 mg/l in the Spoon River. Despite the low amount of sewage in the Spoon River, Kofoid reported that "the organic material in suspension is considerable as indicated by loss of ignition, the albuminoid ammonia, and the organic nitrogen and oxygen consumed." "In the absence of any considerable contamination by sewage it seems probable that these substances have their origin in the organic silt and the soil waters of the very fertile catchment basin of the stream." He refers to the Spoon flood waters as "laden with this organic debris from fertile prairies. The latter thus becomes very important in maintaining the fertility of the river water." Kofoid also reported that "the amount of oxygen consumed is greatest in Spoon River (14.1), and may be attributed largely to the detritus of organic origin which the stream carries, or to the products of its decay held in solution."
Kofoid reports the average concentration of total nitrogen in the Spoon River in 1896-1899 to be 2.59 mg N/l, which is considerably higher than the value of 1.5 mg N/l suggested by USEPA (2000b) for eutrophic waters. The Spoon River was described as largely unpolluted in the 1890s, but could be reasonably described as hypertrophic.
Kofoid reports that "Quiver Lake receives water from the river only during flood periods, when the sewage is diluted, and at other seasons it contains more nearly the chlorine of the uncontaminated prairie stream. Its chlorine thus averages low (3.8)." [Note that the data in Kofoid's table on page 187 shows the average concentration of chlorine in Quiver Lake to be 4.8 mg/l.]
In comparison, Palmer reported the average concentration of total nitrogen in Lake Michigan in 1899-1900 to be about 0.4 mg N/l, which according to the USEPA suggested trophic criteria (USEPA, 2000a) would represent oligotrophic conditions. The average concentration of chlorine in Lake Michigan was reported to be about 3.2 mg/l, about the same as in the Mississippi River and only slightly less than in the Spoon River.
Thus, we see a concentration of chlorine in the Spoon River only slightly higher than in Lake Michigan, but the Spoon River was hypertrophic due mainly to the release of nutrients from its very fertile catchment basin. Other studies show that lakes in Illinois that receive little nitrogen from agricultural lands become hypoxic in summer due to resuspension of high loads of organic nitrogen.
Palmer also reports an average concentration of total nitrogen in the Kankakee River at Wilmington (1896-1900) of 2.86 mg N/l. Average chlorine concentration was reported to be 2.88 mg/l. He reported that "The organic matters contained in the waters of this stream are almost entirely of vegetable origin, for no considerable amount of sewage is discharged into it, that of Kankakee (population 13,995) about 35 miles from the mouth and 25 miles above the point of collection, being the most important." Palmer reported that "there is a considerable diminution in the proportions of nitrates during the warm summer months, this diminution doubtless being in part the result of growth of vegetation in the flowing waters of the stream, in part the result of assimilation of nitrates by the vegetation of the headwaters in the Kankakee marshes, which during this portion of the year constitute the chief source of supply." "The higher nitrates during the high water season are in part due also to the leaching of nitrates from the soil by the run-off and the discharges from the tile drains, which occur chiefly during the seasons of lower temperature and greater precipitation." Again, this is evidence of hypertrophic conditions well before the use of artificial nitrogen fertilizer.
On the mainstem Illinois River, Palmer reported that "..the quantities of organic nitrogen contained in the water of the Illinois River at Kampsville, were in the high water season not less than six and possibly as much as twelve times as great as the quantities contained in the water of the Desplaines at Joliet, which comprises that of the Upper Desplaines, the Chicago Main Drainage Channel or Sanitary Canal, and the Illinois and Michigan Canal." "The much greater quantities of free ammonia [at Kampsville] in March are derived in large part from the rain and the surface wash." "The enormous quantities of nitrates found in the water at Averyville and Kampsville during March and April, the freshet season, are in the main derived from the leaching of surface soils by the run off and the discharge from the tile drains." The great seasonal peaks of all forms of nitrogen in spring is consistent with run-off from the highly fertile and recently cultivated prairies, well before the use of artificial nitrogen fertilizer. The great diminution in the concentration of chlorine at Averyville and Kampsville in spring is further evidence of the non-animal and non-point sources of the spring freshet waters.
Goolsby et al. (1999) report that rates of nitrogen mineralization in soils can be greater than 40,000 kg N/km2/yr in virgin cultivated land. They also report that this mineralization rate in virgin cultivated soils is 3-5 times higher than the mineralization rate measured beneath Illinois soybean and corn crops in recent years (David et al., 1997). Goolsby et al. make no attempt to relate the high mineralization rate in virgin cultivated soils to leaching and run-off. We do.
The National Hypoxia Assessment estimates a pre-development ("pristine") total nitrogen concentration of 1.24 mg N/l for the Mississippi River, without apparently including any enrichment due to regular burning of the prairies. By USEPA criteria (USEPA, 2000b), the pristine Mississippi River would be classified as mesotrophic (0.7-1.5 mg N/l). The Illinois River Basin was and is one of the most organic rich parts of the Mississippi River Basin and the prairies were maintained by burning. It would be reasonable to expect the Illinois River to have a higher-than-average nitrogen concentration in pre-development times. It is reasonable to estimate, using Goolsby's data as a base, that the Illinois River was eutrophic in pre-development times.
We thank you for bringing our attention to Hoskins et al. (1927). We have incorporated the 1921-1922 data for the Lower Illinois River in a revised Figure 30. But we wonder why you failed to draw our attention to some other interesting data in Hoskins et al. on the seasonal variations of nitrogen in Illinois tributaries?
Hoskins et al. data show a very marked seasonal cycle in the concentration of both TN and NO3+NO2-N in the major tributaries to the Illinois River. The concentrations of TN in the Kankakee, Des Plaines, Fox, Vermilion, Mackinaw, and Spoon Rivers in 1921-1922 peaked in December and the monthly average for these tributaries was 6.7 mg N/l. The concentration of TN in all these rivers was lowest in summer (June-August) and averaged 1.54 mg N/l in the lowest months. The concentration of NO3+NO2-N in these rivers peaked in December and averaged 5.14 mg N/l. The concentration of NO3+NO2-N in all these rivers was lowest in August and averaged 0.05 mg N/l. The amplitude of the seasonal cycle of NO3+NO2-N in these tributaries in 1921-1922 was thus considerably larger than the average seasonal amplitude for NO3+NO2-N in all Illinois rivers in 1996, as shown in Figure 15.
Hoskins et al. data also show a pronounced seasonal variation of TN and NO3+NO2-N concentrations in the Lower Illinois River. At the station at river mile 26, the monthly concentration of TN peaked at 5.3 mg N/l in January and reached a monthly minimum of 1.23 mg N/l in June. The monthly concentration of NO3+NO2-N peaked at 3.20 mg N/l in December and reached a monthly minimum of 0.70 mg N/l in June.
The average monthly concentration of TN in the above six tributaries for the 12 months August 1921-July 1922 was 3.6 mg N/l, which was about 23% higher than the concentration in the Illinois River at river mile 26.
The standing N-cycle paradigm cannot explain the magnitude of these early nitrogen
concentrations, or the magnitude of the seasonal nitrogen cycle. Nitrogen fertilizer was not driving the
system in 1921-1922.
THE FERTILIZER HYPOTHESIS.
An important foundation of the fertilizer hypothesis was established in 1970 with the publication of Barry Commoner's seminal paper. Commoner used data produced by the Water Survey to make the point that nitrate concentration in Midwest streams was increasing at the same time that the use of nitrogen fertilizer was increasing, and this was asserted to be a cause-effect relationship. This increase in nitrate from the 1950s to about 1970 is also evident in the water-quality records of other rivers in Illinois, the lower Ohio River, and the Middle Mississippi River. What we present in 2000-8 is a continuation of this time series since 1970. The data show that the concentration of nitrate in the Lower Illinois River peaked around 1970 and subsequently decreased. The concentrations fell as the use of nitrogen fertilizer continued to increase. The strong positive relationship between fertilizer use and nitrate concentrations in Midwest rivers was short lived and broke down.
You now question the quality of the Meredosia data and also suggest that the increase in nitrate concentration could have been due to point-source pollution. Also, you report that the average nitrate+nitrite concentration since 1955 presents a more accurate representation of the concentration measurements in the Lower Illinois than any curve fitting. In so doing, you yourselves are questioning the foundation of the fertilizer cause-effect hypothesis, which was a reported increase in nitrate concentrations caused by nitrogen fertilizer use. If the data are suspect, or the increase in nitrate concentration was due to point source pollution, or there has been no overall increase in nitrate concentration, then this foundation of the fertilizer hypothesis is undermined. Yet you abstain from making the logical conclusion of your own argument and object to our doing so.
You are also critical of us combining data from Meredosia and Valley City to produce a 50-year time series. You note that the nitrogen data at La Grange for 1993-1998 collected by USGS were without width and depth integrated sampling, and were not intended to represent the river as a whole, and should not be used in combination with data from Valley City where depth and width integrated sampling is used. However, data from La Grange are used by USGS to characterize water quality trends in the Illinois River from 1991 to 1997 and LTRMP describes itself as "a river monitoring program" (USGS LTRMP report, 1998). Data from other LTRMP stations are used to characterize water-quality trends throughout the Upper Mississippi River Basin. The above criticism is also hypocritical as one of you (David) in David et al. (1997) reports that "Nitrate values in stream waters over the 6-year period were obtained from a variety of sources, and ranged from excellent to poor in quality." Nitrate records were pieced together from various sources using different sampling techniques sampling frequencies. When direct measurements were not available, seasonal averages from the previous 5-year period were used.
You also suggest that comparison of samples taken in the 1890s and the 1900s is confounded by variation due to sampling location, that N concentration at Valley City would be expected to be lower than at Havana, 60 miles upstream, and that allowance should be made for this in time series analysis. Please note that Goolsby et al. (1999) apparently use historical data from near the mouth of the Illinois River and compare these with recent data from Valley City some 60 miles upstream, although they do not report that the data are from different sites and do not make any adjustments.
Figure 27 in 2000-8 shows nitrate concentrations in the Lower Illinois River since 1975, using data for the time periods for which they are reported in the literature.. The extent to which the smooth curve in Figure 30 fits these data is evident from a visual comparison. We did not attempt any sophisticated statistical analyses or curve fitting for these data. The main point of Figure 30 is to put the 1890's data in the context of 20th century data, i.e. to show that the concentrations of total nitrogen and nitrate increased to a peak around 1970 and subsequently declined. Whatever statistical analysis or curve fitting you apply, this major trend will not change. However, we have refined the diagram and explain it better in the Update. In comparison, Goolsby et al (1999) do not show any data for the Illinois River from the early 20th century to 1980 and, consequently, do not show that nitrate-N and total-N concentrations in the Lower Illinois River peaked around 1970 and subsequently decreased.
I also bring your attention to an article by one of you - David et al. (1997). This is stated to be one of only a few detailed studies "to have linked field N budgets, NO3- loss in tile drained watersheds and surface water NO3- loads." The study characterizes the tile-drained portions of east-central Illinois and the Upper Embarras watershed as homogeneous. From this David et al. apply field nitrogen budgets and averaged NO3- losses in two tile-drained watersheds draining into the Embarras River to determine, among other things, how much of the Embarras NO3- came from these heavily fertilized fields. The data in different parts of the article can be drawn together to show that the nitrate yield from the watershed that received almost 50% more fertilizer than the other watershed had a nitrate yield about 25 % less! But how can this be? This information can be extracted from the tables, but is not discussed in the text. I wonder why not? In an October 15 article in the Decatur Herald, David is quoted as saying that "fertilizer is driving the system." Your own data do not support this assertion - nevertheless, you cling to your belief.
On the same paper, David et al report that "Even if fertilization were reduced or eliminated, the overall disturbance from agricultural production in the Embarras River watershed would still lead to high NO3- concentrations and export, depending on the timing of precipitation events."
The National Hypoxia Assessment reports that the concentration of total nitrogen today is about 40% higher in the Souris-Red-Rainy River system than in the Upper Mississippi River system. Why is this?
I also bring to your attention a recent article by Porter with USGS in the USEPA Technical Guidance Manual for Rivers and Streams (Porter, 2000). This is a study of algal and macroinvertebrate responses to nonpoint source pollution relative to natural factors in the Corn Belt in 1997. Porter concludes that "Nutrient concentrations and the abundance of algae during low-flow conditions were not related directly to rates of fertilizer application or the number of livestock in Midwestern stream basins; however, rates of stream metabolism Pmax and Rmax) increased significantly with indicators of agricultural intensity." Porter finds that algal-nutrient relations were more of a function of landscape characteristics, hydrology, and rainfall-runoff characteristics than agricultural land use, which is relatively homogeneous throughout the region. Porter recommends that "Improved understanding of natural factors and algal-nutrient relations that contribute to chemical and biological indicators of eutrophication in lotic systems could enhance the development of water quality criteria within and among ecoregions in the U.S. (e.g., Level III; Omernik 1986)." You may wish to critically review this report.
The main message remains unchanged. The concentrations of nitrate in the Illinois River, the
Ohio River, and the Middle Mississippi River peaked around 1970 and subsequently declined. In Lake
Decatur, the data show no increase in nitrate concentration since 1967. The statistical and alleged
causal relationship between increasing nitrogen fertilizer use and increasing nitrate concentrations in
surface waters breaks down after about 1970. The almost threefold increase in nitrate concentration in
the Lower Mississippi River from about 1968 to 1983 reported by Goolsby et al (1999) cannot be
attributed to an increase in nitrate concentration in the Midwest rivers.
HIGH NITROGEN FIXATION IN PRE-SETTLEMENT PRAIRIES.
Your position seems to be that Illinois pre-European settlement tallgrass prairie was an undisturbed system. However, it was, as we state clearly, highly disturbed.
Whereas today Illinois is mapped in the Eastern Forest Ecoregion (undisturbed land grows up to trees), 22 million acres of Illinois 35 million acres were in prairie at the time of European settlement. Illinois' prairie was the result of profound Native American land-use practices of regularly firing the landscape. The landscape was further disturbed by the wealth of grazing animals that such land management practices engendered.
The fire-induced prairie acquired about two-fold more nitrogen than it would have under forest. This great accumulation of soil nitrogen occurred in the face of an estimated 90 percent of the above-ground vegetation's nitrogen being lost to the atmosphere with burning (this percentage being greater than crop nitrogen losses due to harvesting) and additional fire-enhanced loss of nitrogen to the hydrosphere and atmosphere by nitrification/denitrification.
From these facts (plus the facts that such burning and grazing pressures greatly enhanced the growth of nitrogen-fixing legumes and nonsymbiotic nitrogen-fixing bacteria) we conclude that nitrogen-fixation in Illinois' pre-European settlement prairies must have been large and likely greater than today's combined natural plus anthropogenic rate of nitrogen-fixation. And, because of this, the then-higher rate of nitrogen-fixation was able to maintain prairie-soil nitrogen levels considerably higher than today's cornbelt soil, in spite of the losses of nitrogen from the prairie landscape.
You state: "First, K & W assume that pre-industrial global N2O and NO fluxes and current anthropogenic N enrichment (specifically fertilizer input) have the same spatial distribution across the globe....By this logic, the highest pre-industrial N2O and NO emission rates must have occurred wherever anthropogenic inputs are currently high. There is no basis for this."
There is no basis for this because we never said or inferred this. This is but a straw man manufactured by you, misattributed to us, and then knocked down by you.
You seem to contend that burning does not increase N2O and NO emissions from soils. And, as with the putative global distribution of natural N2O and NO emissions, you mischaracterize our support for this to chaparrel fire, whereas we used more extensive concepts and data including review of the world literature on the subject. We can add to this to show that fire enhances the emissions oxides of nitrogen to the atmosphere from soils by including the earlier review of Woodmansee and Wallach (1981) who generalize the world literature into their figure 5 which shows the enhanced gaseous N volatilization losses that fire induces through nitrification and denitrification.
To this earlier review we add the later review of Levine et al. (1996) who continue to report that "burning also enhances the biogenic emissions of NO and N2O (Anderson et al., 1988; Levine et al., 1988; Johansson, Rodhe, and Sanhueza, 1988; Levine et al., 1990)."
You state that IPCC (1995) identified oceans and moist tropical forests to be major natural
sources of N2O. This is correct. However, IPCC also states that emissions from biomass burning are
difficult to evaluate. Apparently, IPCC does not even address the emissions of N2O from wetlands. We
address N2O emissions from a regional perspective in the Mississippi River Basin and compare regional
historical direct emissions resulting from burning of the prairies with current regional combustion
emissions. N2O emissions from 110 million acres of wetlands in the Mississippi River basin (Wetland
Matters, 1999) would also have been large and would also have declined with the loss of 66 million
acres of wetlands. In fact, we report on page 47 that "rates of deposition of combustion products
preserved in lake sediments show that for much of the prairie region of the MRB deposition rates
were often tenfold higher in the 0 to 1850 BP period than in the 20th century (Clark, 1996).
IPCC does not address historical regional changes. We attempt to do so. That data from ice cores
show N2O concentrations in the global atmosphere to have increased since the start of the Industrial
Revolution in no way negates our assertions that combustion emissions from burning of the prairies in
pre-European settlement times were high and declined to almost zero by about 1800, and that there
were large emissions of N gases from large areas of wetlands in Illinois in pre-European settlement
times. In the absence of N2O emissions from burning prairies and extensive wetlands in pre-European
settlement times, the global increase in N2O concentration would perhaps have been somewhat greater.
NITROGEN TRANSPORT FROM PRAIRIES TO STREAMS IN SURFACE AND
SUBSURFACE RUNOFF
Undisturbed tallgrass prairies transfer little nitrogen to the hydrosphere, however Illinois' pre-European settlement prairie was not undisturbed.
As previously discussed, the pre-European settlement prairie of Illinois owes its existence to very high levels of disturbance. Whereas today Illinois is mapped as naturally being in the eastern forested ecoregion, west of the Mississippi the vegetation is still mapped as being naturally as prairie. To maintain the "Prairie State" (Illinois) as prairie rather than as forest, disproportionately high rates of disturbance were required to maintain moist tallgrass prairie as the predominant feature of the Illinois landscape.
We extensively document how such disturbances cause large amounts of nitrogen transfers from the geosphere to the hydrosphere. The high nitrogen loadings from Illinois' prairie to its surface waters are further supported by, among other things, the observations of vast areas of aquatic and wetland vegetation, and algae whose nitrogen content equal and exceed those of legumes. The elevated nitrogen fertility of the Illinois prairie elevated the nitrogen fertility of Illinois' water. The Wetlands Initiative cites the loss of the critical nitrogen-removing capacity of wetlands as an important cause of the historical increase in nitrate concentration in the Illinois River from <1.5 mg N/l at the end of the 19th century to average concentrations >5.0 mg N/l in recent years. The Wetlands Initiative also reports that restoring only 407,000 acres of wetlands in Illinois (about 5% of the 1780 wetlands), primarily on flood-prone bottomland throughout the watershed, would remove 101,000 tons (80%) of today's nitrate load from the Illinois River. Over eight million acres of wetlands must have been removing large amounts of nitrogen leaching out of the prairies.
You state that we use Chichester et al (1979) data to illustrate the high rates of nitrogen transferred in surface and subsurface runoff on the naturally-grazed prairie. We did not use these data in this fashion. We used these data "To study the effect of winter versus summer grazing on water quality..." We did not state or imply that the nitrogen runoff rates should be transferred unamended to naturally-grazed prairie. However, we did note that, especially in winter, buffalo congregated along the Illinois River and that animals at all times of the year used surface waters as open sewers much more than today's situation in which animals are generally kept away from open waters. Natural animal congregation by the rivers in winter and in river-side tree shelterbelts in summer, and mud wallowing, are all factors which concentrated nutrients -- especially nitrogen -- thereby increasing the nitrogen saturation of these near water environments.
We extensively document that, while more western prairies are being used as de facto surrogates for nitrogen runoff for pre-European settlement prairies, they are far from being equivalent to the pre-European Illinois moist tallgrass prairie. As one moves from the western prairies into Illinois, the amount of runoff increases, the timing and proportion of runoff moves more into the dormant season, and the landscape becomes more nitrogen-rich. And (unlike the more western locations) moist tallgrass prairie becomes the dominant landscape feature of the landscape in Illinois.
We extensively document that while contemporary studies focus almost exclusively on inorganic nitrogen in runoff (especially NO3-N) the legume-rich, pre-European, moist tallgrass-prairie vegetation loses most of its nitrogen over the dormant season in organic form. And this organic-enriched runoff would occur mostly as surface runoff from moist tallgrass prairie during the dormant season.
We mention that while Illinois is documented as having 2,000 acres of prairie, all of them are affected to some degree by nearby or local drainage. Thus, even if someone was interested in seeing what the runoff chemistry would be, there is no opportunity to do so for what was once the dominant landscape feature of Illinois. Thus we were forced to find a situation that had the most similarity, document the dissimilarities, and then document (at least qualitatively) what the dissimilarities mean.
And so we came up with the Timmons et al (1968) data. We picked the year that had appreciable dormant season runoff (1966 was dry in the dormant season, even by the standards of this relatively dry location). We then noted the dissimilarities between the Timmons et al (1968) experiment and Illinois moist tallgrass prairie: namely, that there would be more nitrogen-rich vegetation to leach and there would be more water to do so in Illinois.
It is this and our comparison of surface runoff to the estimate of current surface runoff nitrogen
for Illinois that you seem to object to. You would, apparently, rather retain western prairie sites as de
facto analogs of Illinois' moist tallgrass prairie, recognize no difference in amount of runoff, timing of
runoff, nature of vegetation (rather have low-nitrogen grasses rather than nitrogen-rich, legume-rich
moist tallgrass prairie), would rather compare surface runoff to estimated surface plus subsurface runoff
(in place of our comparing surface runoff to surface runoff), and do this comparison of the general
situation against "hot spots" of nitrate leaching of your choosing, rather than against state-wide
estimated average.
SCIENTIFIC AND PUBLIC POLICY MERIT OF 2000-08.
In general, the scientific and public policy merit of 2000-08 is to pull together a lot of data and information that exist in the scientific literature in a first attempt to characterize changes in nitrogen richness of the Illinois landscape and waters. This is an important first step. The report puts on the table for open discussion, debate, and further research the following major facts and recognitions:
1. The biggest impact of agriculture on the nitrogen cycle in Illinois has been a large historical decrease in the nitrogen concentration of the soils.
2. Illinois had some 8.2 million acres of wetlands in 1780. Wetlands store, remove, and fix large amounts of nitrogen: 400,000 acres of wetlands will remove 100,000 tons of nitrate, so twenty times this area of wetlands must have been cycling large amounts of nitrogen. The concentration of nitrogen in the surface waters does not represent the amount of nitrogen that is biologically available, but rather the amount of nitrogen left over by the plant and animal production process, denitrification, and burial. All nitrogen cycling needs to be accounted for in conducting nitrogen mass-balance studies and in interpreting historical changes in nitrogen concentrations.
3. The Illinois River and its least-polluted tributaries were hypertrophic in the 1890s. USEPA (2000b) suggests 0.70 mg N/l as the boundary between oligotrophic and mesotrophic river systems, and 1.5 mg N/l as the boundary between mesotrophic and eutrophic systems. The concentration of total nitrogen in the lower Illinois River in 1894-1899 was 3.68 mg N/l, after the assimilation of what must have been large amounts of nitrogen by large expanses of luxuriant vegetation and wetlands, and can be defined as hypertrophic. [Note: the mainstem Mississippi River at Jefferson Barracks had a mean concentration of total nitrogen of 2.7 mg/l in 1899-1900 (Leighton, 1907), only slightly less than the value of about 3.0 mg/l in 1993-1996 at Cape Girardeau (USGS, 1998), and can also be described as highly eutrophic a hundred years ago.]
4. There was a large magnitude to the seasonal cycle of TN and NO3-N in Illinois rivers in the early 19th century.
5. There is a need to address the influence of large herds of bison, elk, deer and other wild animals and birds on the nitrogen content of the landscape and waters in pre-European-settlement times.
6. The pre-European-settlement conditions were not pristine or natural. The prairies were maintained by Native American burning and there is a need to address nitrogen fixation by native tall-grass prairies.
7. There is a need to address the role of extensive fires on the nitrogen content of the landscape and waters.
8. There is a need to understand natural processes as they influence the nitrogen cycle, e.g. drying/wetting and freeze/thaw cycles; temperature and precipitation variations; nitrogen fixation, denitrification, and recycling.
9. The scientific method demands that there is a need to address all forms of human activities as they impact the nitrogen cycle -- losses as well as additions.
10. There is a need to quality assure all data to ensure a sound basis for trend analysis.
11. There is a need to interpret nitrogen concentrations in the context of variations in precipitation and river flow.
12. The concentration of nitrate and total nitrogen in the Lower Illinois River increased from the 1890s to about 1970 and has subsequently decreased.
13. The positive correlation between the increasing use of nitrogen fertilizer and increasing nitrate
concentrations in the Lower Illinois, Middle Mississippi, and Lower Ohio Rivers from the 1950s to
about 1970 breaks down after about 1970.
RESPONSES TO SOME SPECIFIC COMMENTS.
anecdotal. We note that this is also the style of the Wetlands Initiative in describing "Three centuries of environmental history of the Illinois River" (Wetland Matters, 1998).
reasons we do not attach much credibility to this "anthropogenic-only nitrogen" report.
River between the early 1900s and 1980 -- the National Hypoxia Assessment does not.
is made by N fixation!
CONCLUSION
We are proud of our first contribution to identifying reference/background conditions and will continue to provide additional contributions. I will continue to conduct and support work that brings to the forefront of public debate data and information that have apparently been forgotten, ignored, or not used. We hope that, despite your opposition, there will be a lot more discussion and data brought to the table on these important issues.
As far as I am aware, none of you found it important enough to submit detailed comments on the National Hypoxia Assessment reports when they were made available for public review. I can only conclude that, through your silence, you had no problems with these reports from what must have been a multi-million dollar, multi-year assessment, even though the reports fail to use large amounts of relevant data, do not conduct quality assurance on the historical water-quality records prior to 1980, do not attempt to apply flow corrections to these historical data, do not acknowledge depletion of natural nitrogen resources by farming, do not attempt to explain why the prairies are prairies, do not present any data for the Illinois River and Middle Mississippi River between about 1900 and 1980, do not mention that millions of square miles of the world's oceans are naturally hypoxic, do not identify the source(s) of the almost three-fold increase in nitrogen flux from the Lower Mississippi River from about 1968 to 1983 etc. These reports were peer reviewed, approved by representatives of some 20 federal agencies, and finalized before being released for public review and comment. More than 600 pages of public comments and suggestions for improvement were made. It took authors some 9 months to respond to some of these comments, but authors still ignored or chose not to acknowledge or use major comments and additional data provided. The reports remain misleading with extensive distortions and omissions. The October 2000 draft Action Plan, based on these reports, does not provide analysis of social and economic costs and benefits of alternative management plans and, hence, does not meet Congressional requirements. These omissions and distortions apparently do not concern you. Why did you not submit review comments on these important documents, if you are so concerned about the application of science to important policy issues and you know so much about nutrient dynamics? Why do you not hold the National Hypoxia Assessment to the same standards that you try to hold us?
As I testified to Congress in June, we need to i) search out all the available historical water-quality data, ii) assure the quality of the data, with confidence levels, iii) correct the data for variations in rainfall, river flow, plant-life, and other changes in the landscape and aquatic systems that affect nitrogen concentrations, iv) analyze the data for trends, and v) seek to explain these trends. We have a long way to go.
I will place this response and the update on our web site at:
http://www.sws.uiuc.edu/docs/hypoxia/HYPOXIA.asp
REFERENCES
Bellrose, F. C. et al. 1983. The fate of lakes in the Illinois River valley. Illinois Natural History Survey, Biological Notes No.119.
David, MB, and LE Gentry. 2000b. Nitrogen - The state of our State. Illinois Steward 8(4):23-28.
David, M. B. et al. 1997. Nitrogen balance in and export from an agricultural watershed. J. Environ. Qual. 26:1038-1048.
Goolsby et al. 1999. Topic Report 3, "Flux and Sources of Nutrients in the Mississippi-Atchafalaya River Basin." National Science and Technology Council Committee on Environment and Natural Resources, Washington, D.C., May 1999.
Hoskins, J.K., C.C. Ruchhoft, and L.G. Williams. 1927. A study of the pollution and natural purification of the Illinois River. U.S. Public Health Service, Public Health Bull. No. 171.
Illinois Department of Natural Resources. 2000. Illinois Wetlands. Illinois Department of Natural Resources, Office of Resource Conservation and Office of Realty and Environmental Planning, Springfield.
Leighton, M.O. 1907. Pollution of Illinois and Mississippi Rivers by Chicago Sewage. U.S. Dept. Interior, USGS, Water-Supply and Irrigation Paper No.194, Series L, Quality of Water, 20.
Levine et al. 1996. Biomass burning, biogenic soil emissions, and the global nitrogen budget. Biomass Burning and Global Change. Volume 1. Remote Sensing, Modeling and Inventory Development, and Biomass Burning in Africa. J.S. Levine (Ed.). The MIT Press, Cambridge, MA, pp. 370-380.
Mills, H.B., W.C. Starrett, and F.C. Belrose. 1966. Man's effect on the fish and wildlife of the Illinois River. Illinois Natural History Survey, Biological Notes No. 57.
Nelson, R.E. (Editor) 1996. Illinois: a geographical survey. Kendall/Hunt Publishing Co. Dubuque, IA.
Porter, S.D. 2000. Upper Midwest river systems - algal and nutrient conditions in streams and rivers in the upper Midwest region during seasonal low-flow conditions. In USEPA's Nutrient Criteria Technical Guidance Manual: Rivers and streams, EPA-822-B-00-002, July 2000, A-25-A-42.
Talkington. L. M. 1993. The Illinois River: working for our state. Illinois State Water Survey, Misc. Pub. 128, 1993.
USEPA. 2000a. Nutrient Criteria Technical Guidance Manual: Lakes and Reservoirs. EPA-822-B00-001, April 2000.
USEPA. 2000b. Nutrient Criteria Technical Guidance Manual: Rivers and Streams, EPA-822-B-00-002, July 2000.
USGS 1998. Ecological status and trends of the Upper Mississippi River System 1998. USGS La Crosse, WI., LTRMP-99-T001.
Wetland Matters 1998. Three centuries of environmental history of the Illinois River. Vol. 3, No. 1.
Wetland Matters. 1999. Nitrogen farming: harvesting a different crop. Vol. 4, No. 1
Woodmansee, R.G. and L.S. Wallach. 1981. Effects of fire regimes on biogeochemical cycles. Terrestrial Nitrogen Cycles. Processes, Ecosystem Strategies and Management Impacts. F.E. Clark and T. Rosswall (eds.). Ecol. Bull. (Stockholm) 33:649-669.
Sincerely,
Derek Winstanley
Chief, Illinois State Water Survey
Attachment: Update
| cc: | Edward Krug
Brent Manning Joe Hampton Tom Skinner Brian Anderson Don Vonnhame Gary Clarke Jim Park Bruce Yurdin Warren Goetsch Dennis McKenna Kraig Wagernecht Jim Westervelt Susan Adams Jack Erisman David Chicoine Gary Rolfe Bill Shilts George Vander Velde Bruce McMillan Dave Thomas |