David Bouldin, Emeritus Professor, Crop and Soil Science, Cornell University --- E-mail: DRB6@Cornell.edu ---

In 1970 Cornell University received a grant from the Rockefeller Foundation to study runoff from land and its impact on water quality. A multidisciplinary team, focused on the lakes and landscapes in central NY, was developed and led by Professor Robert J Young. Since the above project was finished, I have continued to monitor Fall Creek, sub-watersheds in Fall Creek, other tributaries to Cayuga Lake and the aquifers on the Harford T and R center. (intro.doc)

The appended files describe the results of analysis of over 3000 water samples, 1970- 2003, concerned with land runoff and the lakes in central NY. Major findings follow. References to the appropriate document can be found in the folder "mss"

Three P fractions were measured: MRP, TDP and TP. MRP is measured on filtered samples without treatment and is presumed to be mostly inorganic P in solution. TDP is measured on filtered samples after oxidation of organic forms of P and hence is total Pin solution. TP is particulate P plus TDP. Usually MRP and TDP are considered the major forms used by algae. (ms1, ms2, ms13, ms14)

The average TDP in about 1500 samples from Fall Creek was 0.026 mg per liter, loading was about 4400 Kg P or about 0.13 kg/ha/year . About ? was MRP. Total P was about 0.140 mg/liter. Approximate sources of TDP are as follows: 50% from inactive agriculture and forest, the other 50% attributed to human activities of which about half from diffuse sources and half from point sources. MRP concentrations in runoff from 16 subwatersheds, February to April of 1973, varied from 0.006 to 050 mg/l. The TDP in Cayuga Lake ranges from 0.005 to 0.020 mg per liter (ms3, ms4, ms5, ms6, ms7).

NO3 loading from Fall Creek is about 5 kg/ha /year; this is about 80% of the input of inorganic N in precipitation. This is a consequence of mosaic of sources varying widely in concentration. NO3 loading from 9 subwatersheds in Fall Creek varied from 1 to 7.7 kg/ha/year; no samples containing more than 10 ppm was found. (ms8,ms5,ms9,ms6,ms7,ms10)

Streams draining wooded areas without human habitations or active agriculture have NO3 concentrations and loadings on the order of 20 % of the inputs of inorganic N from precipitation (~1kg/ha/year). and similar to those found in the Catskill and Hubbard Brook in NY (ms5,ms6,ms7,ms9)

There are unlikely to be more than a very few small streams in the Cayuga Lake watershed in which the concentration of NO3-N will exceed the 10 ppm public health standard. However aquifers under heavily fertilized fields may contain more than the public health standard. (ms9, ms5)

Estimates of evapotranspiration (ET) for Fall Creek did not change statistically during the period 1926-1996 as estimated by annual precipitation input minus stream outflow, indicating that land use changes were not important in influencing ET.(ms_15)

The most important sampling protocols are the following: Concentrations of constituents in stream water vary seasonally and with flow intensity. This means that a) timing of sampling must be carried out during all seasons and over all flow regimes, b) amounts of various substances such as N, P and sediment transported to lakes and reservoirs are the product of flow multiplied by concentration which means that flow measurements must be made at the same time as samples are taken for analytical determination and c) most of the water leaves the watershed during the 10 to 20 % of the time that highest flows occur; this means that timing of sampling must include frequent sampling during storm events. With respect to TDP, point sources will be most evident under low flow conditions while non- point sources will be most evident under high flow conditions. Loading of non point sources is thus very much dependent on the 10 % to 20% of the time when highest flow conditions occur. --- The most important conclusion I reached about watershed management is the following. Watershed management requires detailed knowledge about the cost of several management options per unit of decrease in loading/ concentration. Our experience was that the various human activities in sub watersheds were correlated with each other. This meant that statistical analysis of correlations between loading of N and P were useless in identifying the management options which would be most beneficial. This also means that commonly used procedures for validating models are useless in terms of developing management strategies . (Ms12)