Review of the GAPP Science and Implementation Plan (2005)

The committee acknowledges the efforts of GAPP’s Science Advisory Group in developing a science and implementation plan (NOAA-NASA 2004) to further the GAPP science objectives of prediction and decision support. The comments that follow are intended to strengthen the GAPP plan. Since Chapter 1 of the plan consists of introductory material, the committee’s chapter comments on specific chapters begin with Chapter 2 of the plan.

The topic of Chapter 2 is central to the mission of GAPP. The boxed summary in Chapter 2 of the plan clearly and concisely states four memory-related topics as a focus for research to fulfill the GAPP science objective for prediction. The committee endorses the overall summary presented as a solid roadmap for focused progress on this important topic.

Regarding Chapter 2 of the plan and the related Chapter 4 of the GAPP Science Background Document (NOAA 2004), the committee’s major concern is that little is presented in the way of proof of concept of the central promise of improving prediction through incorporating land-surface processes and states. There are no figures, tables, or metrics presented to give a perspective on the state of the art of predictability, the limits of predictability, or the expected geographic areas of enhanced predictability. In a broad sense, the case for the topic is not made, but instead is assumed a priori to be shared by the reader. This is disappointing in that GAPP projects have been funded for four years now by NOAA (see http://www.ogp.noaa.gov/mpe/gapp/abstr/index.html) and NASA (see http://thp.gsfc.nasa.gov); thus, there is potentially a wealth of results published in journals and listed in progress reports that could have been used to define the current state of the art and current research activities and, from these, set up a clear plan for the near future. The lack of synthesis of available results also biases the state of the field by neglecting to mention any negative or mixed results on the topic (contrast, for example, Paegle et al. 1996 with Beljaars et al. 1996, or Findell and Eltahir 1997 with Salvucci et al. 2002). In a broader context, neither the plan nor the science background document mentions or refers to the NRC’s Report of a Workshop on Predictability & Limits-To-Prediction in Hydrologic Systems (NRC 2002), despite its obvious relevance to the topic at hand.

In summary, the plan does not synthesize recent work in the area (including what has been learned from this work and how that has influenced the proposed priorities) or even list what the current and recently supported GAPP projects have been (e.g., those funded by NOAA’s Office of Global Programs and NASA). That context would add a tremendous amount of credibility to the plan for the future. The following paragraphs provide some specific comments and suggestions for improvement. Furthermore, the plan lacks a mechanism for assessing and evaluating the progress of GAPP-funded independent research by the principal investigators.

In Chapter 2 of the plan, the overall objective is well stated and is in fact central to the GAPP mission. One broad-scale issue that is missing is the relation, with respect to implementation, between this chapter and Chapter 6, “Operational Seasonal Climate Prediction: Components from GAPP.” The committee speculates that the distinction, with respect to implementation, is that Chapter 2 is more of a call for studies from principal investigators (PIs), whereas Chapter 6 is more of a guide for the Core Project. If this is the case, it should be made clear.

In the GAPP science background document (NOAA 2004), topography is discussed as having a specific relationship to predictability through processes such as indirect effects on convection and mesoscale circulation. In the plan, however, topography is essentially dropped from this sense (it is specifically discounted as not having memory). Throughout the rest of the GAPP science background document, topography is stressed as important for accuracy in spatial downscaling of predictions. Unfortunately, this distinction is not made in the GAPP plan. The plan should be clear with respect to the distinction between predictability (long memory issues) and prediction accuracy, and it should indicate which of these (or both) is a priority with respect to topography.

In Section 2.1.1 (NOAA-NASA 2004, pp. 9-11) there are four main bullets. The first bullet is presented as quantification of the strength of land-memory processes, but what is described is data analysis. No methodologies are discussed or suggested for this topic. Instead, a set of specific data sets are suggested, which may be too proscriptive. The committee wonders why no modeling analysis was suggested here and why advanced multivariate time series statistical methods are not discussed to address this quantification. Additionally, no sense is given of what exactly (in terms of a quantifiable metric) is even meant by the “strength” of land memory (is it just a decay timescale of soil moisture, or something more subtle?).

The second bullet (NOAA-NASA 2004, p. 10) of the science background section describes an interesting issue related to the spatial and temporal characteristics of memory, but again is not clear as to a metric. A long list of activities is given whose relation to the primary goal is unclear. For example, why does the need to learn about the spatial and temporal characteristics of memory “consequently” involve acquiring or collecting new data for the improvement or validation of models? What are the current shortcomings that require this activity? Strengths here include attempting to use remote sensing observations, since these by definition have good space-time coverage, and testing model abilities for lateral processes, which determines the ability to predict space-time memory (e.g., lowlands have more persistent wetness).

The third bullet (NOAA-NASA 2004, pp. 10-11) of the science background section should provide a diagnostic measure for predictability and, specifically, a metric for improvement of predictability when land-memory processes are included. A significant

amount of research into this issue has been conducted over the past decade (e.g., recent work by Dirmeyer 2000, Schlosser and Milly 2002, Douville 2003, and Koster and Suarez 2003), but no account of what was learned from that work is provided.

The fourth bullet on p. 11 of the plan is well posed. Section 2.2 (NOAA-NASA 2004, p. 11), “Implementation Activities,” claims to use the current status of understanding to define the implementation priority for GAPP research. The section falls short of this goal. No citations are presented to back up strong statements. For example, it is stated that “it was this lack of relevant, regional scale observations of soil moisture that stimulated the land data assimilation system (LDAS) initiative under GCIP (see Mitchell et al., 2004). That successful LDAS methodology will also be adopted and applied in the context of GAPP.” By what measure was this methodology successful?

There needs to be recognition of the advantages of multimodel and/or model-independent approaches to analysis and forecasts of the land-climate system to avoid the biases specific to any single model. Multimodel integrations almost always provide improved results over integrations with a single model—an improvement with no additional model development.

A strength of this section is working with HYDROS (Hydrosphere State Mission) towards measured moisture. Priority activities are listed, but they are not prioritized in the context of their importance or the ability to implement them. For example, the third bullet on improved observations of precipitation (NOAA-NASA 2004, p. 12) is important, but it could be a $10 million plan. The cost of an idea has to play into its priority relative to other potential tasks.

Soil moisture memory issues related to predictability should recognize and acknowledge irrigated agriculture as a large source of variability in moisture, especially in arid areas.

Under Section 2.2.2 (NOAA-NASA 2004, p. 13, “Snow extent amount and duration memory,” the first bullet is well justified. The second bullet on development of submodels for LDAS is not justified as a priority. There have been decades of research into snow modeling, but no sense is given here as to what evidence we have that the current models are inadequate.

The discussion of vegetation and land cover memory in Section 2.2.3 (NOAA-NASA 2004, pp. 13-14) may be an important topic, but as in other parts of this plan, its potential importance is not justified. The plan states that GCIP research demonstrated that “including improved representation of vegetation, even in a simple form, can result in a significant improvement in the predictive capability of seasonal prediction models.” The committee does not know the work to which this refers. How significant was the improvement? It is the committee’s belief that the topic is potentially a good one, but it is hard to sense its priority without knowing how important it is, relative to other issues, in improving predictability.

• The plan should describe and summarize the state of the science for land-memory and variability research, providing particular emphasis on the contribution of GAPP projects.

• The plan should clearly describe the distinction between the focus in Chapter 2 and the guidance in Chapter 6 for the Core Project.

• In the section “Quantification of the strength of land memory processes,” the authors of the plan should consider incorporating modeling analysis and multivariate time series statistical methods to address this issue.

• The plan should provide a more specific, quantifiable metric to measure the strength of land memory.

• The plan should provide a diagnostic measure for predictability, and specifically a metric for the improvement of predictability when land-memory processes are included.

• In addition to simply listing the priorities, the plan should rank the list in order of the importance of each item relative to the availability of funding.

• Plan activities that focus on soil moisture memory issues related to predictability should recognize and acknowledge irrigated agriculture as a large source of variability in moisture, especially in arid areas.

• The plan should provide justification for both “submodels for LDAS” and “vegetation and land cover memory” as priorities.

Chapter 3 begins by emphasizing that GAPP’s expanded focus relative to GCIP outside the Mississippi Basin offers the opportunity for research on the relationship between topography, climatic, and hydrological variables. The chapter is organized into three subsections: “Background” (science and objectives), “Implementation activities,” and “Deliverables.”

In the section on science background (NOAA-NASA 2004, pp. 15-17), the authors provide an overview of the motivations underlying this program element: to predict the spatial distribution of precipitation over the complex terrain of the western United States, to forecast precipitation for intermediate (7 to 15 days) to long-range (monthly-to-seasonal) time scales, and to interpret and transfer the results of improved seasonal predictions for water resource management. The main knowledge gaps identified include the ability to establish theoretical and empirical limits for seasonal forecasts in the western United States; the evolution of clouds and precipitation in mountainous terrain, including the cloud microphysical processes, orographic flow, and thermodynamic effects of mountains; numerical methods; and remote effects involving the space-time evolution of mesoscale systems.

One of the issues raised by the NRC review in 1998 was the limitations of GCIP in addressing the hydroclimatological characteristics of the western United States, especially the impact of mountainous regions. GAPP is to be commended for expanding its geographic region to include the semiarid and mountainous West. The committee agrees that the study of the relationship between topography, climate, and hydrologic variability is an important issue as one of the science goals of GAPP. However, the review of the state of the science and the assessment of science readiness on elements of strategic interest to GAPP is not

comprehensive, while being prescriptive at times, and some of GAPP’s own supported research is not referenced (NOAA-NASA 2004, pp. 15-17).

The second background section (NOAA-NASA 2004, pp. 17-19), “Objectives,” identifies and briefly discusses five research areas to which GAPP is expected to contribute. These include the following:

1. Diagnostic analysis of existing hydrometeorological data

2. Investigation of the mean seasonal cycle and the predictability of its variability

3. Integration of atmospheric elements with surface hydrologic components

4. Hydrometeorological model experiments

5. Application of climate and hydrologic models in seasonal climate forecasting for basins with complex terrain

These activities are appropriate for general research on the hydrometeorology of orographic systems. Yet, as described, they lack the specificity necessary to identify GAPP’s own contribution to the science compared to ongoing research supported by other CCSP partners (e.g., the National Science Foundation, NASA). Furthermore, it would be useful to understand whether and how climate-scale research in the context of GAPP is leveraged by and adds value to other programs at the national level with shared research objectives, such as the U.S. Weather Research Program.

The section “Implementation activities” (NOAA-NASA 2004, pp. 19-21) revisits the previous five areas and adds a sixth linkage to other GAPP elements. For each area, the section lists numerous desirable science activities that follow from the corresponding subsection on objectives, but it lacks an implementation plan with clearly defined priorities, milestones, and outcomes. It is difficult for the committee to comment about how the science objectives will be achieved in view of the plan’s lack of specificity.

Furthermore, a critical issue related to this research priority is the requirement for adequate space-time observations for diagnostic studies, data assimilation, and evaluation of hydrometeorological experiments. Because these observations involve both ground-based in situ and satellite-based platforms at disparate resolutions and characterized by different errors, the development of consistent data sets raises challenges that require extensive scaling and uncertainty analysis. It is thus critical for the GAPP program to describe what arrangements are in place to ensure that these data needs are met.

Finally, the section on deliverables (NOAA-NASA 2004, p. 21) is short, restating in bullet form the six focus areas, along with the expectation and promise that GAPP will indeed fill the existing knowledge gaps and water resources needs. The listed deliverables are too general; and hence it is difficult to assess what the research outcomes will be, and how they will be unique GAPP contributions. As an example, one of the deliverables identified is “improved understanding of the dynamic and thermodynamic effects of mountains,” which by itself describes an entire discipline within the atmospheric sciences. It is worth noting that research related to this issue has been an ongoing effort by many groups for decades. Therefore, what specific contribution will GAPP deliver related to this topic? Similar arguments can be made about the other listed deliverables.

• The implementation strategy and the deliverables of Chapter 3 of the plan should be reexamined.

• It is important that a more comprehensive and inclusive review of the work on this topic be conducted. The specific contributions of GAPP in this area should be articulated in the plan in the context of other ongoing work.

The study of warm season precipitation builds on the lessons learned from GCIP and aims to extend that work to problems related to the understanding and prediction of precipitation in the complex terrain west of the Mississippi River Basin during the warm season. This research element is intended to contribute to both observational climatic data and theoretical understanding of the dynamics of the North American continental warm season precipitation regimes (the so-called North American Monsoon System [NAMS]), which has been initiated by the CLIVAR program. The North American Monsoon Experiment is one of the first GEWEX-CLIVAR coordinated research efforts and is considered a major milestone in bridging two important aspects of climate research. It is also an international collaboration involving both Latin American and U.S. scientists.

The study of the NAMS should provide opportunities to build and expand upon GAPP research strengths with regard to the roles of land-atmosphere interactions, including the dynamical feedbacks associated with snow cover and vegetation heterogeneity, and the role of topography in the organization of convective activity, which modulates the spatial characteristics of precipitation from diurnal-to-interannual time scales. This research focus also provides an opportunity to expand the range of temporal scales in the study of land-memory mechanisms (NOAA-NASA 2004, Chapter 2) and to integrate the study of the hydrometeorology of orographic systems (NOAA-NASA 2004, Chapter 3) into a comprehensive investigation of the predictability of warm season precipitation in the West and Southwest regions of the continental United States.

The background section on objectives (NOAA-NASA 2004, p. 24) identifies and briefly discusses the goals for this research element, which include

1. improved simulation of convective precipitation in climate models;

2. improvements in sustained observations, derived products, and information dissemination; and

3. improvement in operational climate prediction of warm season precipitation.

These objectives are stated without establishing clear linkages to the previous science background section, and they do not appear to be directly related to these goals. The first objective relates solely to the development of improved convective parameterization schemes, but the basis for this is not transparent. Although it refers to GAPP-funded research that has illustrated the sensitivity of model simulations of precipitation to the choice

of convective parameterization schemes in climate and weather prediction models, it goes on to say that this goal should not be a GAPP research activity per se, but should be viewed rather as an opportunity for engaging the GAPP community with climate modelers. This is a commendable overarching goal that should permeate all GAPP- and indeed all CCSP-related activities, but it does not justify the research objective as stated. On the other hand, the representation of orographic effects and land-atmosphere interactions in climate models, which are appropriate GAPP-specific activities, is not mentioned. The plan also does not explicitly state that errors in convective parameterization are a major obstacle to achieving improved forecast skill, but no systematic program has been funded to address this.

The second objective is limited to improved precipitation measurements, but it does not mention any other land-surface observations that would help in understanding the contribution of land-surface processes to rainfall, nor does it mention the problems specific to the measurement of precipitation, soil moisture, surface latent and sensible heat fluxes, and generally boundary-layer processes in regions of complex terrain, which are the unique research foci of GAPP in the context of CCSP.

The third objective relates to improved “operational” seasonal prediction of warm season precipitation. While it suggests that empirical techniques may be a pathway to produce such improved forecasts, it raises the issue of nonstationarity as a possible complication but does not clearly offer a strategy to address this challenge.

Chapter 4 does not contain a specific section on implementation activities; thus, Sections 4.2 and 4.3, which draw on the NAME science plan, appear to constitute the effective GAPP plan element for warm season precipitation. Specifically, the data and activities surrounding NAME are cited as the vehicle for furthering the third objective. On the one hand, PACS and GAPP should be congratulated for pooling their resources. On the other hand, the NAME component of the GAPP program appears to be identical to the NAME component stated in the PACS program. One would have hoped that in this joint PACS-GAPP effort, GAPP would have been able to point to observations, diagnostics, and modeling studies that are complementary to NAME. In particular, one would have hoped for a GAPP-led component of the program that highlighted a systematic investigation of land-surface processes as related to prediction on the monthly and seasonal temporal scales as well as some assessment of the relative importance of land (local) and ocean (remote) boundary forcing. A formally affiliated research project within NAME is the Project to Intercompare Land-Surface Parameterization Schemes San Pedro-Sevilleta land-surface modeling experiment, which is focused on semiarid regions of southern Arizona and New Mexico under the auspices of the GEWEX Land-Atmosphere System Study. This GEWEX connection is not mentioned at all.

The ties to NAME activities are restated briefly in Section 4.4 on deliverables, without identifying specific contributions consistent with GAPP’s science agenda in the context of CCSP. The committee commends the spirit of Section 4.5 on linkages to other programs, although the section could benefit from a diagram illustrating the strategic location of GAPP in the web of existing research programs, with a clear identification of the two-way flows of science and data exchange among CCSP partners.

The research objectives and implementation activities in Chapter 4 should be revisited with the objective of identifying GAPP’s activities and contributions to this research element, thus clearly defining GAPP’s strategic role in CPPA within NOAA and within the broader CCSP research portfolio. In particular, one area that requires further consideration is the balance between terrestrial (land-surface) and atmospheric processes in the context of the need for long-term studies of monsoon systems and the role of orography and land-atmosphere interactions in warm season precipitation.

This chapter of the plan describes GAPP’s relationships with the World Climate Research Programme CEOP activity. It includes a brief summary of CEOP goals and GAPP activities that relate to these goals.

Chapter 5 outlines an impressive set of activities, but it does not clearly differentiate GAPP activities that contribute to CEOP and visa versa. These should be reworded, and the relationship between GAPP and CEOP should be clarified. Other relationships exist, particularly within GEWEX, to elements of the GEWEX Modeling and Prediction Panel and the GEWEX Radiation Panel. Although the overall focus on CEOP is appropriate, especially given the existing close relationship between GAPP and CEOP, the preamble of the document might mention that in the future, collaboration and synergy could be developed with COPES-WCRP, CLIVAR’s Working Group on Seasonal to Interannnual Predictions, and the Earth System Science’s Global Waters Systems Project. The committee’s specific comments below are aimed at improving the chapter.

The major sections of the chapter, 5.1.1 (NOAA-NASA 2004, p. 30), 5.2 (NOAA-NASA 2004, p. 31) and 5.2.3 (NOAA-NASA 2004, p. 34), start with the words “CEOP is,” “CEOP has,” and “CEOP will.” The authors of the plan should rewrite the beginning of each section so that this chapter reads like part of a GAPP plan rather than something extracted from a CEOP plan. The title of the chapter may confuse the relationship between GAPP and CEOP. It implies that CEOP is a part of GAPP. The committee suggests that the chapter title be changed to something like “Contributions to CEOP” or “Relationships to CEOP.”

There are references to GAPP and CEOP global regions, suggesting that GAPP includes studies in regions outside the Americas. The wording should be changed to something like “GAPP and other CEOP global regions.”

The second objective—to demonstrate the utility of operational and next-generation experimental satellites—should be reworded to use the term “evaluate” instead of “demonstrate.” The third objective (NOAA-NASA 2004, p. 31) is a central GAPP activity related to CEOP. The link between this objective and the research activities should be made more explicit in Section 5.2.2 (NOAA-NASA 2004, pp. 33-34).

The relation between GAPP and CEOP’s two international working groups—Water and Energy Simulation and Prediction and Monsoon Studies—is not clearly developed.

According to the plan, “NAME will be GAPP’s major contribution.” This suggests that GAPP has few other major contributions to CEOP. On the one hand, Section 5.2.2 of the plan lists a number of GAPP global products (e.g., global reanalysis and regional reanalysis products). The plan should more clearly identify its contributions of data to CEOP and to what extent, if any, it is involved in CEOP data distribution. For example, the activities listed in Section 5.2 to “develop…data products” read more like a CEOP implementation plan than one for GAPP. This should be clarified.

The oral presentations by the GAPP Science Advisory Group revealed some excellent planned GAPP research activities related to CEOP, but these are not well articulated in the plan. Many of the following comments are suggestions to help improve presentation of that material in the plan.

The lead sentence of Section 5.2.3 (research activities, pp. 34-35 in NOAA-NASA 2004) begins with, “CEOP will increase the value of GAPP (and GEWEX) data sets.” It then goes on to suggest the benefits of CEOP to GAPP. This section would be more effective if it led with a statement of GAPP activities supporting Objective 3.

Section 5.2.3 does not clearly summarize planned GAPP support of research activities for the intercomparison of several models over the NAME region. It is also not clear whether GAPP research would support model intercomparisons for other regions in the Americas (e.g., GCIP, the Large Scale Biosphere Atmosphere Experiment in Amazonia, the Mackenzie GEWEX Study). Section 5.2.3 does not clearly state whether GAPP research would support intercomparisons of the same models over some or all of the CEOP regions including those outside the Americas.

Section 5.2.1 on in situ data and the discussion on deliverables does not identify where these sites are or whether they were established as part of GAPP research or at the request of CEOP, or both.

In Section 5.2 (NOAA-NASA 2004, p. 32), the end of the top paragraph refers to specific information about GAPP contributions to the CEOP remote sensing algorithm development and application. This “specific information” could not be found in Section 8.3 (p. 61). In fact, the second paragraph in Section 8.3 addresses “anticipated GAPP activities undertaken in cooperation with CEOP.” The paragraph does not clarify whether these anticipated activities have already been undertaken or are to be undertaken in the future.

The discussion of deliverables identifies contributions from GAPP to CEOP but does not indicate how these deliverables relate to GAPP data management activities. The plan should indicate whether the data provided to CEOP are identical to the GAPP data sets discussed in Chapter 9 (GAPP Data Management) and should describe any differences.

There have been a number of comparisons between NCEP/National Center for Atmospheric Research, NCEP/U.S. Department of Energy, and NASA’s Global Modeling and Assimilation Office reanalysis (not to mention the European Centre for Medium-Range Weather Forecasts Reanalysis 15 and 40). There is no indication of how these comparisons relate to the GAPP science objectives of prediction and decision support. A discussion of these comparisons should be included in the plan.

• In the plan, the relationships between GAPP and CEOP have to be articulated more clearly to reflect how the GAPP objectives relate to CEOP and how those objectives might be achieved.

• The relationship between GAPP data and related activities described in the GAPP data management plan (NOAA-NASA 2004, Chapter 9) and CEOP data management activities should be documented in the plan.

This chapter is focused primarily on the modeling activities of the Core Project and in many ways is the strongest chapter of the plan. It describes an integrated seasonal prediction system, which is a solid conceptual framework for organizing predictability research for GAPP. However, it is not clear in the plan whether GAPP is proposing an actual system to reside at some major modeling center or a construct to share model-based research. In the oral presentation from the authors of the plan, it was clear that GAPP was suggesting the latter. The plan would be improved if this were more clearly articulated in Section 6.1.1.

The proposed integrated seasonal prediction system is for building and testing a numerical model-based intraseasonal-to-interannual prediction system that includes (1) low-resolution, global coupled atmosphere-land-ocean models to produce sea surface temperature forecasts that are then used to (2) drive a global coupled atmosphere-land model which in turn (3) drives a regional-scale coupled atmosphere-land model to produce forecasts of precipitation and other parameters that (4) serve as input to uncoupled land-hydrology models for runoff, snow pack, and soil moisture. The plan suggests that it will work with “operational” forecast groups to build on existing components of this system. The plan also seems to suggest that it will deliver the missing elements of the entire system to these operational forecast groups. However, it is not clear how this would be implemented. Although a couple of centers are mentioned (e.g., Experimental Climate Prediction Center, International Research Institute for Climate Prediction [IRI]), there is no definition of operational forecast groups. There are no indications whether these or other centers have been contacted or consulted with respect to this aspect of the plan.

The specific objectives of the plan suggest that GAPP will “develop, improve, demonstrate and transition to operations” (1) a global and regional LDAS, (2) improved regional and global land-surface models (LSMs), (3) high-resolution regional climate models, and (4) high resolution hydrology models. This broad charge would be a challenge for the Geophysical Fluid Dynamics Laboratory, NCEP, or the National Center for Atmospheric Research, and seems to be far in excess of the modest resources available to GAPP. Some focus is clearly needed, and specific milestones and deliverables should be identified.

The plan suggests that each of the forecast system components can be developed and tested independently and that some have already been through this process. It is not

clear whether GAPP intends to develop and test new models or to link existing model components together. This would require close collaboration with one of the operational centers discussed above. The NOAA GAPP Core Project resides at NCEP; so as a practical matter, the question is whether adoption of the proposed forecast system is consistent with plans of the NCEP Environmental Modeling Center.

The plan mentions forecasts for monthly, intraseasonal, and seasonal temporal scales, but it is not clear whether the proposed forecast system will address all of these scales. Are the modeling and evaluation studies under the integrated seasonal prediction system framework strictly focused on seasonal-to-interannual precipitation forecasts, or are other temporal scales and variables included?

The plan lacks justification for the potential of hydrologic seasonal forecasting in the United States and the economic value of this forecasting. The committee has already stated in this report that such potential and economic value exist; however, justification should be provided in the plan.

The bulk of the activities and deliverables described in this chapter logically reside within the Core Project. However, during the oral presentations by members of the GAPP Science Advisory Group, there was a clear expectation of community-wide participation in many of the activities. Some indication of the expected contributions to this activity, outside of the work performed by Core Project, would be a welcome addition to the plan.

Experience with atmospheric, ocean, and coupled models suggests that the modeling challenges for climate systems are often fundamentally different from those developed for weather prediction. During the presentation of this chapter at the committee’s public meeting, it was suggested that extending short-term weather prediction models could have value in longer-term predictions. The committee has concerns about this approach and suggests that the authors of the plan evaluate this premise further before including it in the research agenda.

The deliverables listed are extremely ambitious, and it will be difficult to quantify success relative to them. Some less ambitious but more easily quantifiable goals are suggested.

• The plan should clearly articulate whether GAPP is proposing an actual integrated seasonal prediction system to reside at some major modeling center or a construct to shape model-based research in Section 6.1.1. Also, the plan should clearly explain how this will be implemented.

• The plan should provide justification for the potential of hydrologic seasonal forecasting in the United States and its economic value.

• Given the modest resources available to GAPP, the plan should clearly focus and define how GAPP will “develop, improve, demonstrate and transition to operations” (1) a global and regional LDAS, (2) improved regional and global LSMs, (3) high-resolution regional climate models, and (4) high-resolution hydrology models.

• The plan should clearly justify how the proposed forecast system will address monthly, intraseasonal, and seasonal temporal scales.

• The plan should define which deliverables relate to the Core Project and which are expected to result from individual research projects.

• The plan should provide more easily quantifiable goals for the deliverable items listed.

Chapter 7 of the plan begins by reemphasizing the GAPP science objective of decision support. Subsequently, the chapter is organized into three subsections: “Background” (science and objectives), “Implementation activities,” and “Deliverables.”

In the section on science background, the plan authors state that the challenge of accomplishing the decision support GAPP objective is to enable water managers “to understand the nature of probabilistic forecasts and adapt management decision making for such information.” They then introduce a conceptual “end-to-end” process (NOAA-NASA 2004, Figure 7.1) of cascaded elements including climate prediction, hydrologic prediction, decision support tools, and finally, decisions. This section also highlights two important concerns related to the (1) assimilation of observations and (2) handling of input and output uncertainties. With the exception of mentioning macroscale hydrological models, this section does not provide any further overview of or references to relevant GAPP accomplishments.

The second background section on objectives identifies and briefly discusses three priority areas to which GAPP is expected to contribute:

1. Hydrologic prediction and predictability: Determine the key factors governing the ability to predict streamflow, evapotranspiration, and soil moisture and the extent to which incorporation of improved process understanding into hydrologic models results in more accurate hydroclimatic predictions.

2. Hydroclimatic forecasting technologies: Infuse new technologies into hydroclimatic forecasting systems, and assess their potential for both short-term and long-term improvements in operational forecasting.

3. Water resource decision-making: Find avenues to transfer the science contribution of GAPP to the operational hydrology and water resources communities, through demonstration of end-to-end seasonal hydrologic prediction and the use of probabilistic hydrologic forecasts in decision making.

The section on implementation activities begins with the recognition that previous GCIP and GAPP research on water resources focused on individual projects, without programmatic integration of the individual efforts. Subsequently, the section identifies and expands on the following implementation priorities: (1) fostering integrated GAPP activities, (2) community efforts in forecast technologies, (3) collaborative linkages with the operational forecasting community, and (4) collaborative linkages in water resources applications. The activities associated with these implementation priorities are intended to encourage

integration not only across program elements (science and applications) but also between GAPP investigators and related non-GAPP applications-oriented programs. Potential programs and services with which GAPP collaboration opportunities would be explored include the National Weather Service Office of Hydrologic Development and River Forecast Centers, Advanced Hydrologic Prediction Service, Hydrological Ensemble Prediction Experiment, and the RISAs.

Finally, the section on deliverables is fairly general and short, stating that (1) in the short to medium term, collaborations will be established with the National Weather Service Office of Hydrologic Development, River Forecast Centers, and selected water management partners to demonstrate the value of seasonal predictions for water resources; and (2) in the long term, GAPP activities will lead to improved hydrologic forecasts the value of which will be demonstrated in end-to-end applications.

GCIP and GAPP have been pioneering programs in recognizing the need to integrate advances in climate, hydrology, and water resources, within both the science and within the institutional implementation context. Both programs have made significant contributions in this respect by demonstrating the value of integrated science, establishing strong linkages among the climate and hydrologic communities, and developing products with potential value to NOAA’s operational services.

Based on the plan and testimony presented to the committee, the GAPP program Science Advisory Group and program managers wish to demonstrate and promote the value and use of integrated science products to a broader range of water resources users. This is clearly evidenced by the prominent role of decision support in the GAPP stated mission (NOAA-NASA 2004, Chapter 1). This aspect, however, is not well addressed in the plan, most likely because of limited resources. However, if decision support is an important GAPP mission as stated, the committee recommends that this component be considered in the comprehensive manner it deserves. Some specific suggestions follow.

Figure 7.1 of the plan should be expanded to recognize the critical needs for (1) hydrologic as well as consistent water resources demand predictions, (2) decision support systems that can incorporate probabilistic or ensemble prediction products and generate adaptive decision policies, (3) participatory institutional structures that enhance decision making processes, and (4) continuous assessment of the integrated prediction-decision process. See Figure 3-1 below.

The comprehensive consideration of the integrated prediction-decision process would significantly strengthen the prediction and decision support components of the GAPP objectives. Feedback mechanisms from water resources stakeholders have to be developed so that prediction science can better meet their needs. In this respect, it would be useful to expand objective 3 of Section 7.1.2 to individually address decision support system technologies and institutional decision processes:

• Water resources decision support systems: Assess the degree to which current water resources decision support technologies are compatible with new science advances in forecasting, identify key factors for the successful integration of prediction and decision systems, and evaluate their potential operational value.

• Institutional decision processes: Assess and demonstrate the ways that institutional structures enhance or impair decision-making processes.

FIGURE 3-1 Assessment of Integrative Prediction-Decision Process. SOURCE: Aris P. Georgakakos, Georgia Institute of Technology.

These objectives should also be reflected in Section 7.2.1 (NOAA-NASA 2004, p. 52) by outlining more specific implementation activities, and in Section 7.3 (NOAA-NASA 2004, p. 55).

GAPP would benefit from partnerships with agencies that have management and assessment missions, including, among others, the U.S. Geological Survey (especially the Water Resources Institute Program), U.S. Army Corps of Engineers, U.S. Department of Agriculture, U.S. Bureau of Reclamation, Federal Emergency Management Agency, Environmental Protection Agency, and corresponding regional or local agencies. Interagency partnerships could be established within the context of demonstration projects in which participating agencies cost-share resources and jointly develop and/or assess the value of integrated prediction-decision processes and tools. The activities of the IRI (http://iri.columbia.edu) and the Integrated Forecast and Reservoir Management project for Northern California (Georgakakos et al. 2005) are examples of such partnerships and collaborations. Demonstration projects should address a broad range of water resources applications and spatiotemporal scales.

The committee strongly supports the intent to facilitate broad access to retrospective ensemble-type climate prediction products over the continental United States. Based on GAPP’s management and Science Advisory Group testimony, these products will be web accessible and will include multiple traces, daily to monthly time resolutions, and lead times of up to a year. Such products will promote the participation of non-GAPP researchers and will foster community efforts not only in prediction technologies (NOAA-NASA 2004,

Section 7.2.2) but also in integrated prediction-decision technologies. This item should be mentioned explicitly in the discussion of deliverables (NOAA-NASA 2004, Section 7.3).

Section 7.2 describes the intent to establish collaborative linkages in water resources applications to facilitate transfer of improved seasonal predictions into water management operations. It is suggested that GAPP build on the linkages to users established by the RISA programs. Although this is an excellent idea, it does not, in itself, constitute a comprehensive plan to establish adequate linkages with water managers, especially because the RISA program is currently underfunded. Thus, the knowledge transfer plan is not well developed, and the impact of program activities on decision support may not be fully achievable. There is need for a clear action plan to engage the broader water resources research and professional communities (e.g., water supply, reservoir management, hydropower, agriculture, environmental sustainability).

Lastly, the committee recognizes that the transfer of science products to decision makers is a complex undertaking that has challenged and continues to challenge the science and professional communities for several decades (Jacobs and Pulwarty 2003). As stated in a recent NRC (2004a) report, this challenge is exacerbated by the following factors:

• The type and quantity of research needed are unlikely to be adequate if no action is taken at the federal level.

• Responsibility for water resources research has devolved to the states, resulting in the neglect of long-term water resources basic research.

• No structure is in place in the United States to take advantage of the research agendas of identified programs or water management groups.

The committee realizes that GAPP, with its modest budget, cannot support many meaningful activities in this area. It is therefore important that GAPP’s decision support program be designed more strategically and that NOAA and NASA’s application program pay special attention to this important area. CCSP’s water cycle goals (goals 5.4 and 5.5) will never be met unless serious attention is paid to the development of a comprehensive approach to link GAPP and other science programs to applications. There is thus a critical need to augment federal agency involvement and investment in water resources and decision support research and technology transfer. A positive step that NOAA-GAPP can take in this respect is to reaffirm its commitment to the RISA program, secure additional new resources for it, and seek to expand its partnership with other federal and state agencies. The committee also strongly recommends that GAPP explore and benefit from water resources applications experience and opportunities through the GEWEX WRAP program.

• The plan should establish a programmatic plan to bridge predictions with decisions, to understand the context within which decisions are made, and to address the needs of end users and decision makers. Thus, either the decision support objective should be amended or a comprehensive implementation plan should be articulated to support this objective. The committee encourages the latter approach considering that GAPP, by its integrative

mission, is uniquely positioned among federal programs to make pioneering contributions toward the development and application of end-to-end information systems.

• The plan should identify ways to fund and build a more robust and strategic decision support program, including the following:

  • Integrated process aspects: The plan should address (1) consistent climate-hydrologic and demand predictions, (2) integrated and adaptive decision support systems able to explicitly account for system uncertainty, (3) participatory institutional structures that enhance decision making processes, and (4) continuous assessment of the prediction-decision processes.

  • Partnerships: In view of the broad context of the prediction-decision process and the limited resources, the plan should develop a strategy to partner with other federal agencies that work with a broad array of end users and have management missions. Such agencies may include, among others, the U.S. Geological Survey, U.S. Army Corps of Engineers, U.S. Department of Agriculture, U.S. Bureau of Reclamation, and the Federal Emergency Management Agency. The plan should also seek to leverage and strengthen interaction with the RISA program, which facilitates relationships with key users of GAPP science. Because GAPP is in a unique position to support connections with decision makers, the plan should develop an interface with CCSP to accomplish decision support objectives. Lastly, the plan should provide an outline to expand the interaction with the GEWEX WRAP program.

  • Demonstration projects and tools: The plan should support demonstration projects illustrating the value of end-to-end information systems and establishing two-way information flows between the research community and end users. Such demonstration projects should include the active participation of operational agencies and aim to develop climate prediction-decision tools supporting specific applications such as drought preparedness, flood forecasting and management, power planning, and multipurpose river basin management. These demonstration projects should clearly support the plan’s stated objective: “To develop application products for resource managers by interpreting and transferring the results of improved climate predictions for the optimal management of water resources.”

Chapter 8 provides a broad overview of recent advances and future initiatives for satellite-based remote sensing estimates of land-surface states (e.g., moisture and temperature) and forcings (e.g., precipitation). The committee agrees that remote sensing has a clear role in making the measurements necessary to specify parameters and initialize and update coupled climate models, thus helping fulfill the main objective of GAPP. The importance of remote sensing for GAPP helps shape NASA’s potential contributions.

The committee finds, however, that as a plan the main sections of the chapter lack specificity and prioritization, likely because there is no specific office in the NASA research

program where the necessary synthesis of remote sensing data is done. There is research on specific instrument data and assimilation studies with products from individual sensors, but there is no four-dimensional data assimilation of the full suite of satellite data at specified weather-climate prediction centers as noted in Section 8.4 of the plan (NOAA-NASA 2004, p. 61). This is a significant missing function that GAPP alone cannot fill.

The objectives include a wide range of activities, all of which are important in their own right, but only a few of which could realistically be claimed to be of priority to reaching the major objectives of GAPP (seasonal-to-interannual prediction and decision support). The committee finds the third bullet (providing model state constraints) and perhaps the first (providing parameters) to be directly relevant. Two of the bullets (comparison of satellite data with field experiments and validating model outputs) appear to be more tangentially related to the main GAPP objectives (despite, of course, being important activities for NASA to support). The committee recommends that the plan could tie these two issues in through remote sensing activities associated with the NAME experiment. The final issue of understanding the water cycle is too general to claim a direct mapping onto the GAPP objectives.

The section on implementation activities is largely a literature review of past, current, and future hydrologically and climatically relevant remote sensing practices. The activities proposed for vertical fluxes (precipitation and evaporation) largely come down to data set compilation and integration. These activities are not specifically tied to the needs of the Core Project activities of Chapter 6 (e.g., with respect to what variables are needed, on what space and time scales, exactly how they would be used) or to goals of the project that individual PIs might target. The section on horizontal fluxes is incomplete (two short paragraphs) and suffers from issues similar to the above.

Section 8.3 on CEOP is vague. Most of the 10 listed GAPP remote sensing activities are not actually remote sensing activities (e.g., item 2, long-term land model baseline intercomparisons, item 4, land initialization for seasonal-to-interannual coupled prediction, item 5, evaluation of numerical weather and climate predictions for land, or item 9, the production of model output location time series). A few are more related (e.g., item 6), and these should be better described. This entire section needs to be refocused on remote sensing.

Consistent with the science background and implementation activities, the deliverables need specification with respect to the goals of GAPP. Deliverables 1 and 2, if focused on remote sensing, could be useful. As they stand, they are more general data rectification and data assimilation tasks. They do not follow from the rest of the chapter. Deliverable 3 (interconnection of disparate research teams) is not a deliverable, although a plan to accomplish this could be included and is not specific to remote sensing.

• The plan should tie “comparison of satellite data with field experiments” and “validating model outputs” to GAPP through remote sensing activities associated with the NAME experiment.

• Section 8.3 on CEOP (NOAA-NASA 2004) should be refocused on remote sensing.

• In addition to clarifying the role of remote sensing in the first GAPP objective of prediction, the plan should also expand the goal of Chapter 8 to include addressing the second objective of GAPP (i.e., a deliverable on showing satellite data value with respect to water resources decisions—flood management, drought management, agricultural planning, etc.).

• The plan should encourage NASA’s applications program to work with GAPP to demonstrate the value of remote sensing to water management.

The data management plan for GAPP is rather minimal, but already implemented. There are three main components to data management in GAPP:

1. an online catalog of gapp-produced and gapp-relevant data sets,

2. a standardized questionnaire to data set providers to serve as basic documentation for the data sets, and

3. a data management committee that meets periodically at yearly pi meetings and irregularly by telephone to ensure that the data holdings are up-to-date. the acceptance of new data sets may also include some quality control to ensure readability and usability.

This represents the minimal level of data consolidation as defined in the Integrated Geosphere-Biosphere Programme/Biosphere Aspects of the Hydrological Cycle synthesis document (Dirmeyer 2004). That is, disparate data sets from multiple sources have been gathered, virtually, at a single location on the Internet for easy centralized access. This likely represents the greatest return possible on a small investment. Further consolidation (e.g., co-registration of data sets, standardization of formatting or a central data server or browser with search capabilities) requires a much higher degree of technological investment and infrastructure to develop and maintain. Given the limited resources available, the data gathering effort probably represents the best that can be done. The routine and ancillary meetings of the Data Management Committee seem adequate to ensure that current data are made accessible at minimal cost to the program.

Also, there appears to be good coordination with broader GEWEX data efforts (namely, CEOP—although CEOP is not mentioned by name in this chapter), affiliated projects (e.g., NAME), and a statement (in Section 2.1.1, but not in Chapter 9) to adhere to other GEWEX data standards (namely, Assistance for Land-Surface Modeling Activities) as much as possible.

Given the limited resources available and the low probability of increased support, there is little that can be done to improve upon the current plan. However, more certainly could be done with a larger dedicated funding commitment to data management and maintenance for GAPP.

• This chapter of the plan should tie in better to Chapter 5 on CEOP.

• Although GAPP has made great contributions to data management with limited resources, the plan should focus on monitoring for opportunities to enhance data management.

GAPP has a well-defined role as a Continental-Scale Experiment within GEWEX, to which it responds directly through the GEWEX Hydrometeorology Panel on at least an annual basis.

GAPP was until recently a uniquely defined program in the NOAA Office of Global Programs; it has recently been merged with CLIVAR-PACS to become a part of CPPA. This is the main source of support for research and program management, with a dedicated NOAA program manager and additional support from the NASA Terrestrial Hydrology Program. As a result of its major funding role, NOAA exercises significant control over the program.

Within GAPP, the project management structure consists of a Science Advisory Group to provide scientific oversight, a standing committee on data management, and a number of working groups composed of the appropriate interested principal investigators. The working groups have responsibility for some 14 “synthesis products” organized under the GAPP science questions. The committee is concerned that there may not be sufficient representation on the Science Advisory Group from independent scientists (i.e., not current or past GAPP-GCIP PIs).

Beyond the strong affiliation of GAPP with the CLIVAR-PACS program through NAME and the formalization of that relationship in the programmatic reorganization at NOAA, which has led to the formation of CPPA, there is no formal structure in place to interface with research efforts outside GEWEX nor other involved agencies beyond NOAA and NASA. This is partially a reflection of the lack of a coordination structure, such as a U.S. GEWEX project office.

The GAPP synthesis products are largely a restatement of the list of goals, with no associated metrics for measuring progress, and no definition of completion (although a broad time window is given for each and start dates are not identified). This problem reflects back on the lack of a cogent implementation roadmap for the science questions and goals put forward early in the document. Although it is not indicated in the plan, the committee understands that a synthesis document will be produced to record progress in each of these areas. However, this PI-led distributed documentation effort of the synthesis elements is not a substitution for overarching science management.

The committee finds that the organization of management and participants is satisfactory and appropriate, given the nature of the program, its support, and its role within the larger structure of GEWEX and agency support from NOAA and NASA. Nevertheless, the lack of a U.S. GEWEX project office, or even of GAPP project officers for science and

applications independent of the government program managers, may be a major contributing factor to many of the shortcomings outlined throughout this review.

• Membership of the GAPP Science Advisory Group should include persons who have no current direct participation in GAPP science activities as well as those who represent the water resources community.

• The synthesis products in the plan should be better described in terms of metrics or specifications of terms for completion and an actual roadmap for implementation of each.

• Each synthesis product should result in a document or publication specifically describing and quantifying GAPP-supported progress in the area and, where appropriate, its beneficial impact on end users, target communities, or other interested parties.

• Serious consideration should be given to enhanced scientific and applications project coordination and management for GAPP, or more broadly for GEWEX in the United States, that is independent of the government program managers, along the lines of what is already in place for U.S. CLIVAR.