3
Hydrologic Measurements and Observations: An Assessment of Needs

Eric F. Wood

Department of Civil Engineering

Princeton University

Introduction

No one believes a model save its developer. Everyone believes a data set except its collector.—Anonymous

Hydrology is a science built on observations and measurements. Hydrologic theories either have emerged from insights gained through analyzing data or have been confirmed through data that support the theory. The report Opportunities in the Hydrologic Sciences (NRC, 1991) recognizes the importance of data collection, distribution, and analysis by devoting a chapter to the issues concerned with this critical topic. The chapter presents compelling arguments for (1) the need to continue to utilize observational networks and experimental measurements, (2) an assessment of the status of hydrologic data collection, and (3) exploiting opportunities to improve hydrologic data. It seems redundant to repeat the arguments here. Nonetheless, it is possible to provide an assessment of whether concerns raised in the 1991 report have been heeded and whether opportunities have been seized that could provide new, innovative measurements for hydrological theory.

The introduction to that chapter states that:

. . . today there is a schism between data collectors and analysts. The pioneers of modern hydrology were active observers and measurers, yet now designing and executing data collection programs, as distinct from experiments carried out in a field setting with a specific research question in mind, are too often



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 67
--> 3 Hydrologic Measurements and Observations: An Assessment of Needs Eric F. Wood Department of Civil Engineering Princeton University Introduction No one believes a model save its developer. Everyone believes a data set except its collector.—Anonymous Hydrology is a science built on observations and measurements. Hydrologic theories either have emerged from insights gained through analyzing data or have been confirmed through data that support the theory. The report Opportunities in the Hydrologic Sciences (NRC, 1991) recognizes the importance of data collection, distribution, and analysis by devoting a chapter to the issues concerned with this critical topic. The chapter presents compelling arguments for (1) the need to continue to utilize observational networks and experimental measurements, (2) an assessment of the status of hydrologic data collection, and (3) exploiting opportunities to improve hydrologic data. It seems redundant to repeat the arguments here. Nonetheless, it is possible to provide an assessment of whether concerns raised in the 1991 report have been heeded and whether opportunities have been seized that could provide new, innovative measurements for hydrological theory. The introduction to that chapter states that: . . . today there is a schism between data collectors and analysts. The pioneers of modern hydrology were active observers and measurers, yet now designing and executing data collection programs, as distinct from experiments carried out in a field setting with a specific research question in mind, are too often

OCR for page 67
--> viewed as mundane or routine. It is therefore difficult for agencies and individuals to be doggedly persistent about the continuity of high-quality hydrologic data sets. . . . The scientific community tends to allow data collection programs to erode. How is it that observations and measurements have become the stepchild of the science? Even in the chapter titled "Critical and Emerging Areas," there is a preference toward modeling and theoretical developments over data analysis and testing of current theory. This review will look at three related questions in the area of measurements and observations: (1) Has Opportunities in the Hydrologic Sciences had an impact in the area of data collection? (2) Are there areas in which data collection has fallen down? (3) Are there areas where data needs are not being met? The review will be from the author's perspective and for the most part will draw examples from terrestrial hydrology. Data Collection and Operational Hydrology In the United States, data collection in support of operational hydrology and water resources goes back more than 100 years. The needs can be described by considering the aggregated water balance equation over a time interval ?t: (1) where S is the change in soil moisture, P is precipitation, E is evapotranspiration, and Q is runoff. In operational hydrology a major concern is flood forecasting. For floods large enough to put the public at risk, the soil tends to be saturated, so ?S is rather minor and E is negligible, which then leaves the observational requirement of the accurate estimate of precipitation and the timing of the discharge Q. This observational need has had two important manifestations: (1) the development of operational networks operated by the U.S. Geological Survey (USGS) (for streamflow) and the National Oceanic and Atmospheric Administration (NOAA) (for precipitation) and (2) disinterest in establishing observational networks for soil moisture and evaporation by these agencies. Flooding in the United States remains a serious problem that results in significant loss of life and much damage. Figure 1 illustrates that these events are frequent and widespread across the nation. Irrespective of this, there appears to be limited effort by the USGS and the National Weather Service (NWS) to demonstrate, on an economic and social basis, the benefits of an extensive and accurate observational system in support of flood forecasting and warning and to show the relationships among observations, forecasts and warnings, and reduced flood damages and less loss of life. The National Research Council recently assessed the hydrological operations and services provided by the NWS (NRC, 1996), and Mason and Yorke (1997)

OCR for page 67
--> Figure 1 Notable floods and flash floods in the United States, 1987–1991. Source: National Weather Service (1992).

OCR for page 67
--> have summarized streamflow gaging by the USGS. Without belaboring the point, these assessments support what hydrologists have known for some time: observational networks have degraded over the past 10 years, and the accuracy of observations can be improved. Some specific examples to illustrate these points follow. Stream Gaging In 1996 the USGS gaged streamflow at approximately 7,000 stations, a decrease of 185 sites from the previous year and 363 from 1990, indicating an acceleration in site closings. On the other hand, about 60 percent of the sites use satellite telemetry to broadcast stages 24 hours a day to such users as the NWS. Such data are vital to the NWS's mission to provide flood forecasts and warnings. Figure 2 illustrates trends in the USGS's gaging program. What are the strengths? Improvements in the timely delivery of flood stage information can have significant economic benefits, estimated in the February 1996 Willamette Valley, Oregon, floods to be $2.7 billion (Mason and Yorke, 1997). What are the weaknesses? In 1996 the USGS streamflow-gaging program cost the nation $82 million and involved a partnership between the USGS and more than 700 federal, state, and local agencies. While the USGS views the partnerships as a strength, they leave the program vulnerable to the whims of funding from agencies that often view paying for data collection as secondary to other missions in times of shrinking resources. An example is the removal of the gage on the Licking River in McKinneysburg, Kentucky (near Falmouth) in September 1994, which hindered the delivery of accurate flood stage information and contributed to increased loss of life and damages during the March 1997 floods. Neither this gage or its closing was particularly unique. The McKinneysburg gage, used by the NWS for flood forecasting, had been in operation for more than 50 years. It was one of about 15 gages discontinued in 1994 out of a network of 100 USGS gages in Kentucky. As with almost all of the gages in the network of 7,000 USGS streamflow-gaging stations nationwide, this gage was funded as a partnership of multiple agencies` and was operated by the USGS. In this case, 25 percent of the costs were funded by the USGS, 25 percent by the Kentucky Division of Water, and 50 percent by the U.S. Army Corps of Engineers. Network priorities are regularly reevaluated as available funding and priorities change. In 1994 the agencies that supported the McKinneysburg gage determined that other gages were more critical to the missions of the funding partners and that funding would be withdrawn. The USGS and the NWS sought, to no avail, other funding sources to continue operating the gage. Operation of the gage was discontinued in September 1994. As a reference, since 1990 there have been 182 gaging stations closed that were part of the flood-forecasting network of the National Weather Service.

OCR for page 67
--> Figure 2 Streamflow-gaging stations. A, with satellite data collection platforms; B, total number by year. Consider another example of how fiscal pressures are affecting data collection. In Opportunities in the Hydrologic Sciences, the need to develop accurate long-term hydrologic data bases to improve scientific understanding is highlighted. An example is the development of a network of stream gages and water quality sampling stations in undisturbed basins—index stations—that would be benchmarks against which changes in rivers caused by climate change and human development could be evaluated. Such an effort to set up this network is described as ''farsighted" in Opportunities in the Hydrologic Sciences. How has

OCR for page 67
--> the network fared recently? As Mason and Yorke (1997) state: "The most significant reductions [in gaging sites] occurred in the index-stations networks used by the USGS to monitor and document long-term changes and trends in water availability and quality." With these observations, the following recommendation is made: The federal government's requirement for streamflow data needs to be established and a consistent and stable funding mechanism must be put into place. This recommendation is made in light of a 1996 USGS survey of the agency's state-based water resources offices, with input from the NWS river forecast offices, which identified 1,500 sites where additional real-time streamflow gaging would improve flood forecasting and flood warnings. These sites included (1) streamflow gaging stations that were previously discontinued, such as the McKinneysburg gage; (2) locations where new gages could be installed; (3) gages where new measurement devices should be deployed to improve gaging accuracy; and (4) existing gages that need new satellite telemetry to enable them to provide real-time data. The capital costs for the gages are approximately $40 million, annual operation costs would total $15 million, and incorporation of the gages into the NWS flood forecast system would be approximately $15 million. Whether these gages (one or all 1,500) are a good investment for the nation needs to be established within a program that considers the broad national needs for stream gage data. Precipitation In recognition of the importance of precipitation observations in making accurate flood forecasts and timely warnings, the NWS is going through an extensive modernization program. Central to this is the installation of doppler rain radars (WSR-88D). Tremendous strides have been made in observing precipitation systems at fine temporal and spatial scales through the WSR-88D radars. But recent analyses have demonstrated that the resulting quantitative precipitation forecasts are often poor and inconsistent between radars viewing the same precipitation system (Smith et al., 1996a). For example, Figure 3 shows WSR-88D-based estimates of three important quantities—conditional mean rainfall, probability of rainfall, and mean hourly rainfall as a function of radar range—for two radars, Twin Lakes (Okla.) and Tulsa (Okla.). The figures on the left are for the 1994–1995 period, and those on the right are for 1996. Two critical features can be seen. First, there is a persistent range effect on the order of a 50 percent difference between the highest and lowest values, while what is expected are estimates that are constant at all ranges. The algorithmic and radar calibration improvements installed in 1996 did not eliminate this problem. Second, a significant bias exists between the two radars, which are separated by approximately 100 km. Calibration has removed some of this bias, but the difference in mean hourly rainfall is still over 25 percent.

OCR for page 67
--> Figure 3 WSR-88 radar rainfall estimates as a function of range. INX refers to Twin Lakes (Okla.) radar and TLX to Tulsa (Okla.) radar.  The figures on the left are based on 1994–1995 data; on the right, 1996 data.  Source:  Reprinted, with permission from Bauer-Messmer et al. (1997).  © from American Meteorological Society. For individual storms, differences between radars can be unacceptably large. Figure 4 shows total storm rainfall from an intense storm over the Rapidan River basin in West Virginia that occurred on July 17–18, 1996. The radar estimates on the left are from the Davenport, Iowa, radar at a range of about 150 km, while the right side of the figure shows estimates from the Chicago, Ill., radar at a range of

OCR for page 67
--> Figure 4 Comparison of two WRS-88D total storm rain estimates over the Rapidan River (W.Va.) basin. Left, Davenport, Iowa. Right, Chicago Ill. Source: Smith et al. (1996a). © 1996 by American Geophysical Union. about 50 km. The recording gage at the center of the storm reported 440 mm, which is extremely close to the Davenport radar. The Chicago radar underreported the rainfall by about 100 percent. In fact, comparisons between radar-based estimates and gage observations show significant underestimation of heavy rainfall and chronic biases (see Figure 5 from the Tulsa, Okla., radar). The National Research Council assessment of the hydrological and hydro-meteorological services provided by the NWS (NRC, 1996) was quite positive—and rightly so. But what appeared to be missing in the report were quantitative measures and goals in radar calibration, radar-gage comparisons, and forecast accuracy. These measures are critical since they are needed to help allocate funding and to evaluate expenditures among competing needs. This leads to the following recommendation: There is a need to establish quantitative measures and goals in radar calibration, radar-gage comparisons, and their effect on flood forecast accuracy. Operational Modeling Needs Opportunities in the Hydrological Sciences states that ''modeling and data collection are not independent processes. . . . Each drives and directs the other" (NRC, 1991, p. 215). For the NWS, hydrologic water balance models are indispensable tools for producing site-specific flood forecasts and warnings. This is

OCR for page 67
--> Figure 5 Intercomparison of WRS-88D and rain gage rainfall estimates for the Tulsa (Okla.) radar. Source: Smith et al. (1996b). © 1996 by American Geophysical Union.

OCR for page 67
--> more evident as localized flash floods account for a greater portion of flood damages and loss of life, owing to expanding populations and increased development. The components of the surface water balance—that is, the partitioning of precipitation into surface runoff and infiltrated soil water, and the partitioning of the soil water into evaporation and transpiration, drainage, and stream discharge—are influenced by surface characteristics such as soil texture, vegetation, and topography. In the past five years the availability of such data has increased tremendously. For example, 1-km soil texture data are widely available, the USGS earth resources observation system (EROS) data center provides weekly advanced very high resolution radiometer (AVHRR) satellite vegetation data, and digital terrain data are available at 90 m nationally and 30 m for about half the country. The expanding role of operational models for water management and environmental monitoring and prediction has resulted in the need for continuous time (storm and interstorm) modeling of stream discharges and river basin water balances. Considering Eq. (1), this requires that each variable either be specified or represented mathematically based on an understanding of the underlying hydrological processes. This suggests that we need to represent infiltration processes to estimate ΔS/Δt; that evapotranspiration must either be measured and used as an input or estimated from the surface energy balance; and that streamflow, which is made up of surface runoff and drainage from the infiltrated water into the drainage network (baseflow), must be accounted for. For modeling severe flooding, ΔS/Δt and E tend to be negligible, and for the design of water resource projects, climatological estimates of E are sufficient (since design by its very nature represents expected performance). Nonetheless, long-term accurate data sets are needed to determine site-specific constants for operational hydrological models. It is critical that hydrologists continue to articulate this need. In operational weather forecasting models, such as those used by the National Center for Environmental Prediction (NCEP) or the European Centre for Medium-range Weather Forecasts (ECMWF), the needs are somewhat different. In the past five years there has been a drive to incorporate better representation of terrestrial hydrology, which represents the lower boundary condition to the atmosphere. It has been shown that such improvements lead directly to improved weather forecasts (e.g., Beljaars et al., 1996). The Beljaars et al. (1996) paper is a particularly important one because it reported on the influence of land parameterizations on precipitation forecasts from weather prediction models. The installation and testing of cycle 48 (CY48) occurred during June 1993 and paralleled forecasts produced during July 1993—a period of particularly heavy precipitation and flooding in the central United States. The major improvement to the land surface model included a better representation of the upper soil column moisture and heat flux dynamics—based, to a great extent, on data collected

OCR for page 67
--> Figure 6 Total precipitation over the United States for July 1993 (A) and the percentage of normal precipitation (B) as published by the Weekly Weather and Crop Bulletin (August 3, 1993). (A) The contours are at 0.5, 1, 2, 4, and 8 in with light and heavy shading above 4 and 8 in., respectively; (B) the contours are at 25, 50, 100, 200, and 400 percent, with shading above 100 percent. Source: Beljaars et al. (1996). under the First international satellite land surface climatology project (ISLSCP) Field Experiment (FIFE) held in central Kansas during the summers of 1987 and 1989 (Sellers et al., 1990, 1993). Specifically, the operational CY47 system had a land surface scheme that

OCR for page 67
--> Figure 7 Mean forecast precipitation of all 48- to 72-hour forecasts verified between July 9 and 25 with (A) CY47 and with (B) CY48. The contours are at 1, 2, 4, 8, etc., mm/day.  The printed numbers are station observations in millimeters per day.  Source:  Beljaars et al. (1996). was heavily constrained by a climatologically defined deep soil boundary condition for soil moisture and temperature, whereas in CY48 the model produces a dynamic, time-varying, surface soil moisture through time integration of the atmospheric forcings of precipitation and evaporation. As an example of the impact of the modified land surface representation (summarized by Beljaars et

OCR for page 67
--> al., 1996), consider Figures 6 and 7. Figure 6 gives the July 1993 total precipitation over the United States as reported in Weekly Weather and Crop Bulletin. Figure 7 shows the mean forecast precipitation from ECMWF 48-to 72-hour forecasts from July 9 to 25, 1993. These dramatic improvements of precipitation forecasts—especially in the 48-to 72-hour forecasts—have been the foundation of stated desires by the atmospheric community for extensive soil moisture measurements. In a similar manner, there has been a dramatic improvement in surface air temperature forecasts in high latitudes by incorporating improved winter time albedo measured as part of the ISLSCP Boreal Ecosystem Atmospheric Study (BOREAS) (Viterbo, ECMWF, 1997, personal communication). These improvements support the stated need in Opportunities in the Hydrologic Sciences to develop hydrologic data bases to improve scientific understanding (in this case, improvements in representing underlying hydrological processes and the needed parameters in operational models). Over the past decade the climate experiments carried out under the auspices of the World Climate Research Programme (WCRP)—hydrologic-atmospheric pilot experiment (HAPEX)-MOBILY, HAPEX-Sahel, ISLSCP-FIFE, and ISLSCP-BOREAS—have provided important data that have been used to improve operational weather models. There has been a tremendous improvement in available data to evaluate coupled water and energy balance models (e.g., the improvements in ECMWF's model reported by Betts et al., 1993; Viterbo and Beljaars, 1995; and Betts et al., 1996). But measurements at experimental sites did not start with these programs. The Agricultural Research Service (ARS) of the U.S. Department of Agriculture operated, and still operates, a large number of experimental catchments across the United States. While these experimental catchments had agricultural hydrology as their primary research focus, they collected a wealth of data that potentially could be used to improve the land surface hydrology of operational models. For example, over the past few years there has been strong interest in soil moisture measurements because of the coupling between soil moisture and evapotranspiration and soil moisture and infiltration—with their subsequent effect on both the energy and the water balances. Current parameterizations of transpiration in operational models (e.g., ECMWF, see Viterbo and Beljaars, 1995, or NCEP's Eta model, see Chen et al., 1996) include functions (Jarvis, 1976) that limit transpiration due to environmental factors, including soil moisture stress. Yet many ARS catchments have long periods of soil moisture measurements in conjunction with precipitation, evaporation, and meteorological data. Over the past few years, the ARS's hydrology laboratory has attempted to organize some of this data into an on-line database (see http://hydrolab.arsusda.gov). There are other important hydrologic data, some collected by the ARS and other agencies, that are sitting in boxes under desks, in field stations, on PCs, and who knows where. Considering the interest in and importance of historical data for improving hydrology models, the following recommendation is made: An assessment

OCR for page 67
--> should be carried out of what data are needed to improve the representation of hydrological processes in operational models, the availability of such data, and the effort to incorporate critical data sets into an on-line database. Measurements and Scientific Hydrology In the early 1980s there was a recognized need for large-scale climate field experiments to provide the necessary data for achieving progress in (1) coupled land-atmospheric processes for climate modeling, (2) understanding water and energy exchanges at General Circulation Model-grid scales, and (3) utilizing remote sensing data for validating large-scale hydrological models. Opportunities in the Hydrological Sciences recognizes these needs and the importance of the synergism between modeling and measurement: "the development of hydrologic theory [and] data collection are interconnected [and] measurement of hydrologic variables is a scientific endeavor itself" (NRC, 1991, p. 215). This section addresses progress made in the past five years to the identified "opportunities to improve hydrologic data" discussed in Opportunities in the Hydrological Sciences. Coordinated Experiments During the past decade, the field programs developed under the auspices of the WCRP have provided critical data that have advanced scientific hydrology. Recent HAPEX-MOBILY, HAPEX-Sahel, ISLSCP-FIFE, and BOREAS experiments have shown the value of high temporal resolution water and energy flux measurements in evaluating land surface hydrological models (Shao and Henderson-Sellers, 1996; Chen et al., 1996) and making improvements to operational weather prediction models (Betts et al., 1993; Viterbo and Beljaars, 1995; Betts et al., 1996). Carried out in Kansas during the summers of 1987 and 1989, FIFE provided (for the first time) consistent high-quality water and energy flux data suitable for coupled water and energy modeling. This allowed hydrologists to develop and validate coupled water and energy balance models at catchment scales (Famiglietti and Wood, 1994a, b; Liang et al., 1994; Peters-Lidard et al., 1997). The interactions between experimentalists and modelers, among scientists covering terrestrial hydrology, boundary layer processes, biospheric processes, and remote sensing, were unprecedented. The 764 journal pages in the first FIFE special issue (Journal of Geophysical Research , 97(D17), November 30, 1992) are a testament of the level of scientific activity arising from the FIFE data. But the real success of FIFE was National Aeronautics and Space Administration's (NASA) open data policy, in which data are freely available to all from the time of collection. In addition, the effort was made and money was spent to incorporate the data into a database. The availability of these data, first through CDs and

OCR for page 67
--> currently over the Internet, played a critical role in expanding the scientific impact of FIFE beyond its initial scientists. Following NASA's lead, there has been greater openness and availability in distributing data from other WCRP climate studies. Two main goals identified from these coordinated studies, stated in Opportunities in the Hydrologic Sciences, are (1) understanding the energy and water cycles at scales beyond the field plot and (2) combining the satellite and in situ measurements in order to verify large-scale hydrologic models. The observations taken under the WCRP experiments have been of mixed value for large-scale terrestrial hydrology. This is because the major programs had foci that only included terrestrial hydrology peripherally. HAPEX focused on hydrology atmospheric parameterization. ISLSCP was formed in 1993 from interests in remote sensing, climate, and ecological communities to look at the appropriateness of visible/infrared (VIS/IR) satellite data to study land surface biogeophysical properties and their role in partitioning incoming radiation into latent and sensible heat. In a similar manner, the formation by ICSU (International Council of Scientific Unions) of the core project on the Biospheric Aspects of the Hydrological Cycle (BAHC) had as its focus ecological and biophysical terrestrial processes. Only through GEWEX, the Global Energy and Water Cycle Experiment, is there a direct goal linked to terrestrial hydrology. The products of the HAPEX and ISLSCP program goals are data sets of limited usefulness for modeling and validating water and energy balances across a range of scales from field/tower scales to regional and continental scales. The experimental plans emphasized detailed spatial measurements for short periods (which are more suitable for developing satellite retrieval algorithms) over long-term measurements. In addition, there was more emphasis on biophysical and atmospheric measurements than terrestrial hydrologic fluxes. Therefore, carrying out basic analyses such as a water budget during the FIFE experimental period is difficult due to incomplete and uncertain data (Duan et al., 1996). In fact, it is impossible to use data from any climate experiment—including the U.S. Department of Energy's cloud-atmospheric radiation testbed/atmospheric radiation measurement (CART/ARM) experiment—to carry out an observation-only water and energy balance over a catchment or grid of 10,000 km2 or greater. As another example, the BOREAS experimental plan made no allowance for accurate measurement of precipitation, and only through separate funding from the National Science Foundation and Environment Canada was a rain radar system installed for the Southern Study Area. Insufficient stream gaging and inadequate measurements of surface infiltration, interflow, percolation, and ground water movements have resulted in unresolved differences between measurements of water and energy fluxes at different BOREAS scales. It appears that the supporting role of hydrology in these experiments resulted in inadequate planning and funding for critical hydrologic measurements. Over the long term, the hydrological data from these experiments should be the most valuable data. As

OCR for page 67
--> agencies start planning for the new ISLSCP experiment in Amazonia, it is likely that inadequate planning and resources will go into the hydrologic measurements and analyses of hydrologic processes. This deficiency will result in scientists being unable to answer basic questions posed in the scientific plan. This leads to the following recommendation: There is a need to better define the hydrologic data needed to achieve the stated goals of the WCRP climate experiments, including goals related to atmospheric and biophysical processes, and to strengthen the terrestrial hydrological studies within the WCRP, HAPEX, and ISLSCP climate experiments. As an aside, the GEWEX-Mississippi River basin study (GCIP) and GEWEX-Mackenzie have considered the terrestrial hydrologic data needs more carefully. GCIP has performed well in gathering together hydrological data to understand hydrological processes across scales up to the continental watershed scale, and this has resulted in significant improvements in NCEP's operational Eta model. Nonetheless, the program has relied on operational data whose quality may be insufficient for the goals of GEWEX or of scientific hydrology. Such operational data include precipitation from the WSR-88D rain radar, streamflow from the USGS network, surface meteorology from surface airway stations, atmospheric profiles from NOAA's radiosonde network, and radiation from the geostationary GOES satellite. Each of these have shortcomings in reaching the GEWEX goal ''of determining water and energy fluxes by measurements of observable atmospheric and surface properties." As an example, consider Figure 8, which shows accumulated precipitation over a seven-day period in the Red-Arkansas basins. Systematic errors among different WRS-88D radars resulted in identifying different radar footprints. Such errors may have a limited effect on operational forecasts, but they greatly affect the accuracy of long-term water balances. Thus, the following recommendation is proposed: For GEWEX experiments utilizing operational data, there is a need to establish the accuracy requirements for data related to estimating water and energy balances across a range of scales. Remote Sensing Readily available satellite observations represent one of the most positive aspects of the past decade and have been brought about by a combination of WCRP/ICSU initiatives like ISLSCP, NASA's Earth Observing System (EOS) research program, and advances in computer storage and transmission. The challenge here is not acquiring satellite data but rather processing and transforming the data into needed biogeophysical parameters. This last step requires broadened ISLSCP-type programs to acquire ground-truth data for the development of satellite retrieval algorithms, and a set of (global) ground validation sites. NASA currently plans to fund a validation program

OCR for page 67
--> Figure 8 WRS-88D accumulated precipitation estimates over the Red-Arkansas river basins region of GCIP for April 8–14, 1994.

OCR for page 67
--> TABLE 1 EOS Instruments and Products and How They Will Be Used By Land-Surface Climate Models Instrument Products Uses ASTER Land surface temperature, snow cover, cloud characteristics, albedo Visible and near-infrared bands, elevation, albedo Forcing Parameterization Validation CERES Albedo, radiation fluxes, precipitable water, cloud forcing characteristics, surface temperature aerosols, temperature humidity, and pressure profiles albedo Forcing Parameterization Validation MIMR Precipitation, snow cover, soil moisture albedo, clouds fraction, aerosols, soil moisture bidirectional reflectance distribution function Forcing Parameterization Validation MRIS Temperature and water vapor profiles, cloud cover, albedo, surface temperature, snow cover, aerosols, surface resistance/ evapotranspiration land cover classification, vegetation indices, leaf area index/fractional Photo-synthetically Active Radiation Parameterization Validation SAGE III Cloud height, H2O concentration and mixing ratio, temperature and pressure profiles Validation TRMM Rainfall profile, surface precipitation (precipitation radar, TRMM microwave imager, visible/infrared imaging radiometer Forcing, validation Other 4 dimensional data assimilation, including wind, temperature, and humidity profiles; momentum, energy, and precipitation fluxes Forcing, validation   Source: Running et al. (1997). for sensors under EOS, but very little is oriented toward hydrological programs. Why? Because terrestrial hydrological modeling, through remote sensing, will require a suite of satellites while the validation plans tend to be oriented on a single-sensor basis. Table 1 gives a list of EOS instruments that are needed for land surface hydrologic modeling. In the terrestrial ecology community, more progress has been made in establishing a coordinated global set of terrestrial observations (Running et al., 1997). A Terrestrial Observation Panel (TOP) was established under the Global Climate Observing System (GCOS), a body established to provide the observations needed

OCR for page 67
--> to meet the scientific requirements for monitoring, detecting, and predicting climate change. TOP is developing a five-tier terrestrial observation plan and implementation strategy in conjunction with the WCRP and the International Geosphere-Biosphere Programme. The five-tier approach helps to establish the level of financial and scientific activity required for different types of validation efforts. The approach recognizes the necessary participation of large but temporary projects like FIFE, HAPEX, and BOREAS for certain validations and more permanent, geographically distributed facilities like national resource station networks for other global validation activities. This organizing vision is essential to produce globally consistent and representative validations of the full suite of land science products. The potential exists for a similar plan for terrestrial hydrology, but little has been accomplished to date. Thus, the recommendation: There is a need to establish a global validation plan for terrestrial hydrology, and the plan should be an integral component of the GCOS. Summary and Conclusions This paper has attempted to evaluate progress in observations and measurements since the publication of Opportunities in the Hydrologic Sciences. The progress is mixed: On the positive side, there is a substantial body of new measurements, thanks in part to WCRP-coordinated experiments; there is increased awareness and appreciation among hydrologists that measurements and observations are a critical component of developing new hydrologic theory and models; and data availability has increased due to significant advances in computer processing and data storage and transmission. On the negative side, the hydrology community has failed to establish the data needs for either operational hydrology and water management or scientific hydrology. Because of this, terrestrial hydrology has been the ''stepchild" of climate field experiments, resulting in incomplete or inadequate measurements that prevent basic computations like determining water and energy balances over a range of scales. In addition, operational data collection remains under fiscal pressure, resulting in short-term savings at the expense of long-term benefits. Meanwhile, the hydrology community has yet to clearly articulate the operational data needs of the nation. References Baeck. M. L., B. Bauer-Messmer, J. A. Smith, and W. Zhao. 1998. Heavy rainfall: Contrasting two concurrent Great Plains thunderstorms. Weather and Forecasting 12(4):785–798. Bauer-Messmer, B., J. A. Smith, M. L. Baeck, and W. Zhao. 1997. Heavy rainfall: Contrasting two concurrent Great Plains thurderstorms. Weather and Forecasting 12(4):785–798. Beljaars, A. C. M., P. Viterbo, M. Miller, and A. K. Betts. 1996. The anomalous rainfall over the USA during July 1993: sensitivity to land surface parameterization and soil moisture anomalies. Mon. Wea. Rev. 124: 362–383.

OCR for page 67
--> Betts, A. K., J. H. Ball, and A. C. M. Beljaars. 1993. Comparison between the land surface response of the European Centre model and the FIFE-1987 data. Q. J. R. Meteorol. Soc. 119: 975–1001. Betts, A. K., J. H. Ball, A. C. M. Beljaars, M. Miller, and P. Viterbo. 1996. The land surface-atmosphere interaction: A review based on observational and global modeling perspectives. J. Geophys. Res. 101(D3):7209–7225. Chen, F., K. Mitchell, J. Schaake, Y. Xue, H. Pan, V. Koren, Q. Duan, M. Ek, and A. Betts. 1996. Modeling of land surface evaporation by four schemes and comparisons with FIFE observations. J. Geophys. Res. 101(D3):7251–7268. Duan, Q. Y., J. Schaake, and V. Koren. 1996. FIFE 1987 water budget analysis. J. Geophys. Res. 101(D3):7197–7207. Famiglietti, J. S., and E. F. Wood. 1994a. Multi-scale modeling of spatially-variable water and energy balance processes. Water Resour. Res. 30(11):3061–3078. Famiglietti, J. S., and E. F. Wood. 1994b. Application of multi-scale water and energy balance model on a tallgrass prairie. Water Resour. Res. 30(11):3079–3094. Jarvis, P. G. 1976. The interpretations of the variations in leaf water potential and stomatal conductance found in canopies in the field. Roy. Soc. Lond. Philos. Trans. Ser. B 273:593–610. Liang, X., D. P. Lettenmaier, E. F. Wood, and S. J. Burges. 1994. A simple hydrologically based model of land surface water and energy fluxes for general circulation models. J. Geophys. Res. 99(D7):14415–14428. Mason, R., and T. H. Yorke. 1997. Streamflow Information for the Nation. Fact Sheet FS-006-97. Office of Surface Water, U.S. Geological Survey, Reston, Va. National Research Council. 1991. Opportunities in the Hydrologic Sciences. Washington, D.C.: National Academy Press. National Research Council. 1996. Assessment of the Hydrological and Hydrometeorological Operations and Services. Washington, D.C.: National Academy Press. National Weather Service. 1992. Flash Floods and Floods: A Preparedness Guide. NOAA/PA 92050. Washington, D.C.: U.S. Government Printing Office. Peters-Lidard, C., M. Zion, and E. F. Wood. 1997. A soil-vegetation-atmosphere transfer scheme for modeling spatially variable water and energy balance processes. J. Geophys. Res. 102(D4):4303–4324 Running, S., J. Collatz, J. Washburn, and S. Sorooshian (eds.). 1997. EOS Science Implementation Plan, Land Ecosystems and Hydrology. Washington, D.C.: National Aeronautics and Space Administration. Sellers, P, J., F. G. Hall, G. Asrar, D. E. Strebel, and R. E. Murphy. 1990. The First ISLSCP Field Experiment (FIFE). Bull. Am. Meteorol. Soc. 71(10):1429–1447. Sellers, P. J., F. G. Hall, G. Asrar, D. E. Strebel, and R. E. Murphy. 1993. An overview of the First International Satellite Land Surface Climatology Project (ISLSCP) Field Experiment (FIFE). J. Geophys. Res. 97(DI7):18345–18371. Shao, Y., and A. Henderson-Sellers. 1996. Modeling soil moisture: A project for inter-comparison of land surface parameterization schemes phase 2(b). J. Geophys. Res. 101(D3):7227–7250. Smith, J. A., M. L. Baeck, M. Steiner, and A. J. Miller. 1996a. Catastrophic rainfall from an upslope thunderstorm in the Central Appalachians: The Rapidan Storm of June 27, 1995. Water Resour. Res. 32(10):3099–3113. Smith, J. A., D. Seo. M. L. Baeck, and M. D. Hudlow. 1996b. An intercomparison study of NEXRAD precipitation estimates. Water Resour. Res. 32(7):2035–2045. Viterbo, P., and A. Beljaars. 1995. An improved land surface parameterization scheme in the ECMWF model and its validation. J. Climate 8: 2716–2748.