McDonnell’s Ideas on How to Write a Successful NSF or

USDA ProposalIdeas on FE research as we move to OSU 2007

and more integration across campus

Hey John, I wonder if I can get off campus rates on our next NSF proposal

The facts

NSF Hydrological Sciences program receives about 120 proposals per round—of these they fund 10-15!

USDA NRI odds are better—perhaps double

It is a difficult thing to write a proposal that gets funded.

Nevertheless, we must do it!

Like many things, it gets easier with practice and a bit of instruction

Outline of this talk

• How NSF and USDA “work” in terms of their rfp’s, reviews, panels

• How to map your research interests against the program manager’s expectation

• How to structure the proposal

• The most important bits• General philosophy on

curiosity-based research

How I think NSF defines research

Research is to see what

everyone has seen and think

what no one else has thought

Albert Szent Gyorgyi

..

This is different to applied research

How you should view research that you propose to NSF or USDA

Research must be fundable, laudable and achievable. All three together define success.

How the NSF process works

• You submit a proposal in June• It goes out for review June-Aug (5 reviews solicited buy

usually 3 received)• The Panel meets in September—each panelist handles 20 or so

proposals– They come to the meeting to lobby on behalf of a few in

their keep• Panel ranks proposals based on review strengths and their own

perspective• Panel recommends 15-20 proposals and Program Manager

selects top dozen or so• Almost always a budget negotiation with the Program

Administrator• Process repeats itself in December

The USDA NRI Process

Essentially the same process as NSF:• Submit Nov 1 deadline• Out for 5 reviews• Panelist sent 20 or so proposals each to handle• Each panel member “lobbies” for a sub-set of his or her

proposals• Panel meets in late March or early April• Excellent and Very Good necessary for next step• Some discretion among panel members• Ranked list to Program Manager• Telephone call in early May (this week!!! I’m

waiting…)

The Panel

• 2-3 days in Washington

• 10-15 minutes per proposal at the Panel

• Will have 20-25 to shepherd through the system

• Will rely on reviews and the Summary Statement– The most important part of

the proposal!

Example of a (successful) Summary StatementPROJECT SUMMARYAlthough knowledge of the hydrologic flowpaths is critical to the preservation of public water supplies and the understanding of the transport of point and non‑point source pollutants, we have not been able to model the linkages between hydrologic flux and geochemical flux. This is a result of the fact that models and field studies generally are not well linked in hydrological investigations in catchments. Furthermore, attempts to introduce such field observations into models have faced the inevitable compromise between the value of additional data and the identifiability of parameters required to take account of the data. New research on runoff flowpaths has shown that spatially-distributed models (but not conventional lumped approaches) are required to capture the key hydrological and geochemical processes operating in the watershed that affect water quality. Therefore, funding is requested to develop a new synergistically linked hydrologic/geochemical model based on the realization that: (1) waters evolve chemically and isotopically along the different flowpaths in the catchment, and that (2) these specific flowpaths may produce distinctive water "signatures" related to the topographic position and hydrologic history of the water.

Uncertainty analysis procedures will be used as a means to assess uncertainty associated with the model parameters. Our objectives are: (i) to develop a hydrogeochemical model within which the dynamics of stream chemistry and flowpaths can be assessed and directly related to data from the fieldwork campaigns, (ii) to realistically assess model structure/parameter uncertainties and to use this assessment to effectively identify fieldwork strategies to reduce such uncertainties, (iii) to make spatially-distributed measurements in the field to validate spatially-distributed estimates of soil moisture, water table elevations, isotope and solute concentrations, (iv) to examine the controls on geochemical/isotopic evolution along flowpaths during events and on between-storm residence times by using physical characteristics and responses of the catchment, and (v) to use the new model as a tool for testing specific hypotheses of hydrologic flowpaths and mixing in the catchment.

Overall, the expected results and significance are two fold; (i) this study represents an innovative attempt to link a evolutionary reaction-path model constrained by both solute isotope and geochemical data within a topographically-driven hydrologic model developed at the hillslope scale (hence the development of new fieldwork techniques and insight); and (ii) a new type of hydrogeochemical model will be developed, in a form that is constrained by the interaction of the field campaign and the assessment of model uncertainty (hence new calibration strategies, model development techniques and, of course, a new model), that attempts to resolve some of the problems of lumped and fully distributed models and can be readily applied to other study areas.

Example of a (successful) Summary Statement

The Reviewer

• A busy scientist with too many demands on her/his time.

• Will compare yours with the 2 or 3 others that they have been asked to review

• Will read it in 60 min or less• Will compose her review in less

than 30 min

Therefore, the proposal must be extraordinarily well written

NSF is not risk tolerant

• Use seed money to show proof of concept

• Show a key graph or table demonstrating some chance of success

• Go big or stay home….– Challenge some top model or some top

procedures and involve the creators (See Schoenholtz seminar on how to build relationships next)

Within NSF

It’s NOT an old boys club• Big names get rejected just as often as new

comers (I’ve seen this when I have sat on panels)

• Having had an NSF proposal has no bearing on your new proposal– Except that if you repeat the winning

formula, your chances are much higher

…well, not an old boy’s club in that way…most do have white skin, gray hair and male anatomy…with CoF we have VERY successful women role models in the area: Barb Bond; Bev Law and others

The Guts of an NSF Proposal

• You need a new idea • The solution should lead to

new understanding outside the field site per say—transfer value

• You are asking for money—always remember this! Why does your idea merit anyone giving you money?

The Guts of a USDA Proposal

• You need a real problem • The solution should lead

to new understanding outside the field site per say—transfer value

• You are asking for money—always remember this! Why does your idea merit anyone giving you money?

A Quote from an NSF Research Program Manager

• “90% of the grant’s likelihood of success is based on how novel your questions are—ideally they are ones that have not ever been thought of or posed before”.

The idea

• Formulate as null hypotheses that can be rejected if at all possible

• What will be the lasting contribution of the work?– A new model

– A new analysis technique

– A new conceptualization

Photo: G. Grant

The hydrology of the Santiam River watershed

The role of rainfall thresholds for activating hillslope contributions to

Cascade stream channels.

The Proposal Title

Photo: G. Grant

The Scientific INTRODUCTION• Needs a “snappy” lead sentence to catch the reader’s

attention. This is CRITICAL to the proposal.• Includes literature background pertinent to the

PROBLEM, not the field site chosen to address the problem.

• Includes discussion why previous studies have been insufficient to resolve the problem

• Near end indicates why the field site chosen (or experiment etc.) was chosen for the study.

• Ends with specific HYPOTHESES to be tested.• Ease of writing: Very difficult.

From Don Siegel

Key Things for Clarity

Questions (a few)• Hypotheses stemming from the questions

(many)• Methods that relate directly to how

hypotheses will be tested• What will be the contribution: new model,

new understanding (if so, how will it be delivered?)

On references

• If you have good papers on the topic in the top journals, then cite heavily

• If not, do not cite your papers from conf proc and lowly ranked journals—rather cite key papers from others from Forest Science, CJFR, etc.

Don’t sweat the budget

• I would argue that the budget amount and details (unless unusual) do not make much difference

• Average grant size in Hydrological Sciences is $90/yr for 3 years

• Find out what these figures are for the program to which you will apply and get into that ball park

• Put in 1 mo of salary per year even if you are on a 12-mo appointment– Then lobby Steve/Hal to release an equivalent

amount to you to help build your program• USGS, USFS can be co-PI• Indirect on first $25K of subcontract

If you find that a topic just doesn’t want to get funded

• Try something else

• Don’t beat your head against a wall

• There are lots of problems to explore

• Work different angles

• Try another agency

One strategy

• On a grant, create a team to cover the key bases.

• Team up with high fliers.

• Increasingly, multidisciplinary research projects are funded over single disciplinary projects.

Bring together sub-fields not yet integrated

CPlan

AVHRR

DHSVM ln a/tanB

A new approach to quantifying landuse change in watersheds

Propose a new way forward on a problem where field is stuck

• Spectral analysis• Isotope analysis• Analytical Hierarchy

Model• Visualization• New geophysical tool• Data mining• Portable weather radar

Expected Results and Significance

• Demonstrate how you have thought about how your findings will be used

• How findings will influence other fields

• How findings will challenge existing paradigm

• How findings will challenge existing model

Example of a (successful) Expected Results and Significance section

D. Expected Results and SignificanceOur proposed research will result in a spatially distributed model of geochemical flux, driven by topography. The principal benefit of our work to the hydrologic community will be an improved understanding of the linkages between hydrologic flux and geochemical flux in small catchments. For improving hydrological understanding, reactive solute isotopes will provide new insights into flowpaths that are not discernable with δ18O and tensiometry (i.e., identifying flowpaths labelled by chemical reactions). In terms of improving geochemical understanding, the development of a physically realistic hydrologic model structure to approximate flowpaths, semi-distributed to account for spatial variability in water chemistry controlled by residence time (as a function of topographic convergence) and soil properties that evolve along flowpaths (catena), will be a major step forward in understanding catchment geochemistry.We will attempt to quantify the uncertainty in geochemical predictions at the catchment scale and to evaluate the value of different types of data in identification of model parameter value distributions. This is an important "next step" from the optimal parameters developed by Beven and Binley (1992). Secondly, we will use the model as a tool for testing identifiable hypotheses, for example the new water mixing problem of Robson et al. (1992), or the McDonnell (1990) hypothesis that macropore flows are old water. If such a tool can be used and verified at Panola, it would have allow other researchers to determine the minimal data set required apply the model in other watersheds.

Example of a (successful) Expected Results and Significance section ‘cont

Overall, the expected results and significance are two fold: (i) this is the first study to attempt to model evolutionary reactive-solute-isotope data combined with geochemical reaction-path data (within NETPATH) on the hillslope scale (hence the development of new fieldwork techniques and insight), and (ii) a new type of hydrogeochemical model will be developed, in a form that is constrained by the field work and uncertainty principles (hence new calibration strategies, model development techniques and, of course, a new model), that attempts to resolve some of the problems of lumped and fully distributed models and should be readily applied to other study areas. In short, models which are developed interactively with field measurements are very rare. Most modelers stop after fitting the parameters to match reality; they don't USE the model as a tool for developing testable hypotheses, which can then be verified/refuted with field measurements. We plan to make model development and field work inextricably linked throughout the three years of proposed funding.

The Politics of Scientific Proposals

How NSF Proposals are Scored:1. Excellent = Fund it!2. Very good = Fund it if there is money3. Good = Don’t fund it; proposal needs work4. Fair = Proposal is really bad. Furthermore, I

think this researcher is inept. 5. Poor = I personally want to destroy the career

of the person who proposed this shit or she/he is walking on my turf.

From Don Siegel

Once you have the $$ in hand• Leverage, leverage, leverage• Cite it, cite it, cite it

– Papers, presentations, web

• Publish, publish, publish• Develop a rapport with the program

manager• Write related proposals to NSF,

USDA, EPA, BLM, USGS, CALFED, etc

• Submit future NSF proposals that build on the outcome and the format

Conclusions • Go for it, it’s not that bad• NSF and USDA money is prestigious• At most universities it is expected for a

positive tenure decision• It shows that work is peer reviewed up

front…this has a cascading effect on the success link: publish-proposal-grad students; publish-proposal-grad students………

• It will force you into a mode of research that will benefit all that you do—just look at Barb Bond, Bev Law and Mark Harmon!