Nutrients National Synthesis

Ideas for Future Work

Here are brief descriptions of the topics NNS team members think are high priority for NNS and NAWQA that will be the basis of discussion about future NNS work at our team meeting in Sacramento, April 21-22, 2009.

The main points are to:

• identify those items central to finishing Cycle I and II analyses,
• retain some capacity to address issues and perform analyses that would help answer questions relevant to the Cycle III design.

Tom

Scale up of a mechanistic model to support (1) aquifer vulnerability maps that show concentrations as well as probabilities, and (2) (with Larry) long range forecasting for climate change effects on nitrate leaching. Scale up would be accomplished at flowpath and trend sites, which have multiple observations to support calibration by inverse modeling. For (1), the final hybrid model would simulate pesticides as well as nitrate.

You can see the basic idea for (1) in the attached proposal, but there have been some changes to my thinking. That is, focus on trend and flowpath sites to enable more reliable model calibration.

Larry

Tom and I have been discussing a possible future application of the RZWQM calibrations we have just completed for the ACT study. Basically what we have noted is the large amount of nitrogen stored in the organic pool of the various settings. It is massive compared to the amount of N that is in the inorganic pool and what leaches to the water table. Movement of N in and out of the organic pool is regulated by the rates of mineralization and immobilization. These rates are dependent on temperature and soil moisture, both of which may potentially change in the future given the predicted changes in global climate. We propose using calibrated RZWQM scenarios from around the country along with projected changes in temperature and precipitation to predict future changes in the amount of nitrogen mineralization and groundwater recharge, which together allow us to calculate future nitrate recharge concentrations. Next week Tom will give a short briefing on the model calibrations and I will follow with more detail on this proposed work. We think that this study would be of interest to the scientific community as a whole as well as those focused on climate change, the agricultural community, the public policy community, and anyone drinking water from shallow aquifers.

Norm

1. Base flow analysis of ortho P loading
   a. Methods as we used for nitrate.
   b. Include look at running a single LOADEST model. Use all data to calibrate but use different flow values (total and base) for simulation.
   c. Include some additional work on snowmelt sites to determine if refinement to BFI is possible.

2. Trends in low flow / base flow concentrations and variability in low flow / base flow concentrations.
   a. Link with ground water trend work
   b. Incorporate Lehn’s ideas and seek input from Lehn
   c. Include analysis of what constitutes a representative nutrient sample for ecological analysis.
   d. Where is hydrograph separation an adequate tool versus selection of a time period representing low flow / base flow.

3. Occurrence and distribution of sediment
   a. This is one of EPA’s areas of interest.
   b. LOADEST values for analysis
   c. QA/QC analysis of NAWQA sediment data
   d. Comparison of cone split vs. standard methods sediment concentrations and loads. Can we even determine collection methodology from NWIS data coding?
   e. Publish NAWQA sediment data set.

Lori

I attached my earlier email below. Greg and I have recently discussed #1 as a possibility next FY. And I am still particularly interested in #2.

Something else: We always say how critical it is to understand the hydrologic conditions during sampling, as hydrology can complicate spatial evaluations. But I haven't seen hydrology explicitly considered together with other factors (i.e. our map of HIP mean Q/ Wolock's long term mean Q in the circular is just a map of one of numerous factors that might be affecting water quality - it doesn't get into how much influence the hydrology actually had in our spatial evaluations). I'm not sure at this point how to go about this - would need to think on it more. It's possible that Norm's baseflow analysis could tie in with something like this as well.

1. Nonmonotonic long-term trends

Issue: Trend models based on time trend with constant level and slope can be too restrictive in the long term, where changes may be nonmonotonic. In structural time series models, these parameters are essentially assumed to follow random walks, leading to a stochastic trend in which the level and slope are allowed to evolve over time. These models have previously been used in economics and climate, but not hydrology -- the censoring and irregular sampling in most water-quality data sets will require modification of previous methods.

Potential objectives: (1) Develop methods for detection of evolving time trends appropriate for long-term water-quality data sets. (2) Determine nonmonotonic trends in nutrient concentrations for selected long-term monitoring stations. (3) possibly... Compare trends in concentration to timing of major water-quality policy and source changes over the last ~30 years (or however long we decide to look at).

Potential collaborators: Greg Schwarz

Risks: Method development may not pan out, in which case we would fall back on more traditional but more limited long-term trend assessment methods. Would need long-term source data for comparison - not sure what's available at the moment.

2. Climate change

The latest IPCC report on climate change states that "Widespread increases in heavy precipitation events (e.g., above the 95th percentile) have been observed over the late 20th century, even in places where total amounts have decreased.... It is very likely that heavy precipitation events will become more frequent." The latest report from the U.S. Climate Change Science Program also projects that precipitation, on average, is likely to be less frequent but more intense in the future. These findings are leading to speculation that increases in heavy precipitation events have led to (and will continue to lead to) increased non-point source runoff.

Potential objectives: (1) Determine whether the fraction of annual load carried during extreme events has increased in the last ~30 years at selected long-term monitoring stations. (2) Determine whether total load has increased as a result. (3) Evaluate seasonal changes in the timing of extreme events (i.e. are more extreme events occurring during fertilizer application periods?).

Potential collaborators: Not sure yet. Probably folks who could provide consult on statistics, precip and runoff generation, and/or climate change.

Risks: May also need to evaluate trends in total and extreme precipitation and runoff, an area of heavy and active research (and of course debate). Not sure if we'll run into problems with the fact that our long-term data sets are at large sites, where impacts of localized extreme events could be dampened. Will most likely need to normalize loads to factors other than # of extreme events-- including source changes, total precip, etc-- and would need data and methods to do so. Loads will be estimated prior to any subsequent analyses -- need to figure out how to account for estimation/sampling error and biased hypothesis tests that could result when using these estimates.

3. Climate variability

Issue: Recent studies have documented quasiperiodic variations in precipitation, ground-water levels, vadose zone infiltration, and streamflow related to interannual and interdecadal climate forcings. Little work has been done to date on understanding water-quality responses to this climate variability, but it can be hypothesized that changes in precipitation, runoff, ground-water transport, etc might affect nutrient concentrations and loads in streams as well. Understanding the impacts of climate variability could support future water-quality management decisions.

Potential objective: Determine whether changes in nutrient concentrations and loads coincide with known climate cycles such as ENSO, the North Atlantic Oscillation, etc.

Potential collaborators: Mike Dettinger and Randy Hanson developed a method that has been used previously with hydrologic time series. Jason Gurdak (in CO) recently applied it in a study of the unsat zone in the High Plains.

Risks: Method may need to some tweaking for application to concentration/load data.

4. Nutrient criteria exceedances

Issue: States are developing nutrient criteria with data at a subset of their stream reaches, and could benefit from broader insight into where/why exceedances are likely to occur.

Potential objective: (1) Predict the probability of nutrient criteria exceedances in US streams. (2) possibly... Identify factors increasing the probability of exceeding these criteria. (3) Compare results among land uses, political boundaries, climatic regions, etc.

Potential collaborators: Individual state environmental agencies, if we wanted to use criteria they are likely to use (instead, could use EPA recommendations). Could create new logistic regression models, or possibly leverage SPARROW or ROBIN, so Norm, Rich, and others.

Risks: No major methodology risks I can think of at the moment. Except possibly poor point source data.

5. QA/QC of STORET and other data

Issue: NAWQA is putting increasing emphasis on the use of non-NAWQA water-quality data, and little systematic consideration has been given to data quality and comparability. Relatively reasonable for some applications (ie SPARROW), but not others.

Potential objectives: This could involve a wide variety of things... such as units checks, identification of outliers/data entry errors/duplicate data, evaluation of possible site overlaps, examination of major spatial and temporal gaps, co-location of streamgages, etc. Some of this has been done as part of the SPARROW efforts, but only under a limited set of conditions.

Potential collaborators: With the right oversight, many different people could work on this. Ideally someone with database and coding skill.

Risks: Only risk really is not doing this.

Neil's feedback for Lori

Thanks Lori. I haven't really thought about this - I'm not clear on how much "normalization" we achieve by using flow-weighted means, and what additional information we could extract. We might get somewhere by thinking about the practical questions. For example, a question that EPA may ponder could be an amalgum of also all of what you have below: how will transport of nutrient loads to receiving waters be impacted by climate change, and at the other end, how will nutrient conditions critical to aquatic ecosystems respond to climate change (not that I'm hedging my bet here - we would likely assume that this means low flow nutrient concentrations, but what little I know suggests this is an over-simplification)

In fact, I think most of what you have below could be worked into one broad strategy that starts with trying to identify climatic forcing (as they like to put it) in the QW record, and ends with trying to forecast the outcome of change.

I. - -Method development: 1. Nonmonotonic long-term trends.....you may need this to do step two, ... - #5 could be a subset: optimize use of existing data

II. Existing Record: 3. Climate variability....Determine whether changes in nutrient concentrations and loads coincide with known climate cycles (and related Q # 2 (2) below) - could split into high flow and low flow question

III. Prediction a. Impact of change in high flow: How will loads to receiving waters change in different areas based on projected change in flow regime? b. Impact at low end: #4 below rotated so it asks the question for projected conditions at low flow

So, just some reactions on my part.

Other items discussed previously:
- the very narrow question of reliability of representing nutrient conditions by one low-flow sample: where does this work or not?

Karen

Karens's trends proposal and Brant's description of his scope of work.

Mark

Well, I guess I should jump in the conversation. Below are a few ideas for the next couple of years, followed by some more general ideas for short-long term. The audience for these include states, EPA, and USDA.

1. Phosphorous in GW - Based upon the work of Jim Tesoriero NAWQA could do more with the relative contribution of phosphorous from GW to surface waters. Most of the GW/SW interaction work is very focused on nitrogen, which is partially historical and misconceptions. But P is still a major issue and may be a much larger issue than we tend to think. This issue is important in regards to stream recovery and criteria in that if nutrient reduction occurs in a basin, will groundwater provide sufficient nutrients to drive the system production.

2. Flow weighted concentrations - I think someone, and maybe they are, should take a hard look at the differences between flow weighted concentrations and other concentration measures. SPARROW uses loads to determine concentrations, but I have no idea if there is any bias in that approach. Dale Robertson told me he was concerned about this, and I think he may be addressing this, but not sure.

3. Land use and BMPs - I am not referring to percent land use, but to more refined coverages like riparian systems. Could we look at basin-level data on the extent of riparian systems and relate to nutrient loads/concentrations in streams.

4. Single vs multiple nutrient sampling - ENS may have addressed this issue, but we commonly get asked about the comparison of single nutrient samples we use to assess some biological measure. During summer periods single samples seem to work as well as multiple samples, but what about other times of the year.

Future things to think about:

1. Biogeochemical cycling and biological response - There is a need to assess N, P, and C cycling and how this is influenced by stream habitat, sw/gw interactions, and biological systems. This would include sediment chemistry.

2. Real-time continuous monitoring - There is a lot of interest in real-time monitoring of nutrients, turbidity, dissolved oxygen, and other parameters. We tend to focus a lot on variation due to changes in flow, but nutrients change over a 24 hour period. Take a look at the pp by Steve Kalkoff.

3. Climate - There are a lot of concerns about change in climate and how that will influence the interactions of nutrients, primary production, and in-stream communities.

4. Stream process - We need to improve our understanding of nutrient processing in streams in regards to management practices, stream habitat complexity, and in-stream communities.

5. What if models - There is a lot of demand for models that will permit us to make "what if" predictions. If I change A will I see a change in B and C. This could be in terms of nutrients themselves, or ecological effects. SPARROW does this somewhat for loading, but that is all it does.

Bill

Let me add a couple of comments to the question about assisting with the cycle III design. Gary Rowe's team may ask for special assistance from individuals to complete the cycle III plan; but, I don't think we can anticipate these areas at the moment.

The priorities that I've shared with the NLT are as follows. The highest priority is to complete existing synthesis and topical team reports related to cycle II that we have made commitments for. Second priority for the national synthesis teams is to assist with the completion of other existing high priority program goals for Cycle II that are regional or national in scope and technical-support type issues that will put the program in a stronger position to address the program's long standing goals--status, trends, and understanding that will still be a focus for cycle III.

There are several ecological synthesis "understanding" products that I've discussed rescoping with Daren that should include a greater level of involvement from hydrologists and chemists to compliment the efforts of the ecologists. There are several status and trends efforts that may fall into the same category. The desire is that these products should reflect a more complete understanding of how the systems (key landscape factors, hydrologic variables, chemistry, and stream ecology) that we are assessing change with time.

An example of the type of technical support work that synthesis resources will support is a prioritization of trace organic analytes for cycle III that is being led by John Zogorski. Another will be a retrospective (and perhaps a reconnaissance effort) focused on polonium and other naturally occurring radionuclides that have gotten visibility during cycle 2. There are other technical support issues related to pulling together data sets needed for future modeling--(e.g. conservation tillage, Point sources ). Still others could focus on statistical techniques needed to accomplish key objectives for the program.