8. COLLABORATIVE RESEARCH ACTIVITIES

GCIP has evolved from its beginning as largely an international project to a largely national project with participation from many different agencies in the USA. This evolution has fostered the development of cooperative and collaborative activities in many different areas.

8.1 Collaboration with Other GEWEX Projects

Adequate description of hydrologic processes is required in global models of the ocean-atmosphere-land system to improve the prediction of weather and climate at all time scales. Research is required to make best use of the data available from GCIP and other GEWEX large-scale observational programs to guide the formulation and validation of such hydrologic submodels. Improving the description of hydrologic processes in global models is a priority issue for GCIP which will be best addressed in collaboration with PILPS, ISLSCP, and the GEWEX Hydrometeorology Panel.

8.1.1 Research relating to the GEWEX Cloud Systems Study

The goal of the GEWEX Cloud Systems Study (GCSS) is to improve the parameterization of cloud systems in climate and NWP models. This objective will be achieved through a better quantitative knowledge of the physical processes involved in cloud systems as well as a quantification of their large-scale effects ( GCSS 1994). Key issues are described in Browning (1994). The investigation of continental cloud systems is part of the long-term objectives of the GCSS Working Group on Precipitating Convective Cloud Systems (Moncrieff et al. 1997).

One of the aims of GCIP is to improve the treatment of surface and hydrologic processes in NWP and climate models, but clouds have an important impact on these processes. GCSS involvement would contribute to the cloud component to GCIP, by way of cloud-resolving modeling and related activities. In turn, the GCIP data sets would be used to evaluate these models against observations.

Cloud Resolving Models

Cloud resolving models, identified by their ability to resolve cloud dynamics, are the approach of choice of the GCSS. These models derive from traditional nonhydrostatic cloud models but their scope is more ambitious. The effects of convection on the environment and the interaction among physical processes (boundary layer, surface layer, radiation, and microphysics) are the pacing issues, rather than individual processes per se. Since the time scales of some interactions (e.g., cloud--radiation) can be weeks, this is not only demanding on model design but also requires large computer resources.

When used to study precipitating convection (e.g. Grabowski et al. 1996a, b) or frontal cloud systems (Dudhia 1994) grid lengths of about 1km can be successfully employed to calculate bulk effects. Consequently, the domains of cloud resolving models span many NWP grid volumes. The time scales examined by 2D models is up to several weeks and these models are poised to address issues on intraseasonal time scales. An example is the effect of cloud-radiation interactions on the atmospheric and surface energy budgets (Wu et al. 1995b).

Cloud-resolving models also explicitly resolve convection-mean flow interactions that are impossible to accurately observe and since cloud-scale dynamics is explicitly simulated, one key uncertainty is minimized. Data sets from cloud resolving models can be used to evaluate single-column climate models - the testbeds for convective parameterization schemes. These data sets are also a key element in formulating new and more comprehensive approaches to parameterization.

Models need to be evaluated against atmospheric data sets. The GCIP region features several cloud system types, ranging from deep precipitating convection during the warm season, to frontal clouds dominated by ice processes in winter. GCIP will provide data sets for evaluating cloud resolving models, noting the relatively high density of routine observations over the U.S., not to say the special long-term observations available from the ARM/CART site.

Two different types of evaluation are required. First, an evaluation of the physical parameterizations used in cloud-resolving models (e.g., microphysics, turbulence, surface processes and radiation) is needed. However, this requires detailed cloud-scale observations, as well as intensive observation periods involving airborne platforms. Neither is available from GCIP.

Second, the effect of clouds on the environment directly relates to convective parameterizations in GCMs and is, in principle, an area to which GCIP can contribute. It is, however, far from a simple matter to utilize data collected during the GCIP Enhanced Seasonal Observing Periods (ESOPs) to evaluate the models.

A basic issue is: what is the minimum observational detail required to evaluate cloud resolving models? An ultimate answer will involve data assimilation in both regional and global models to "fill in" missing or data-void areas. However, present assimilation methods are neither a panacea nor even practicable on cloud resolving model grids. GCSS will therefore focus on basic problems such as the ensemble response of clouds (deep and shallow) to spatially-averaged, time-dependent forcing applied over scales comparable to or exceeding, climate model grid scales.

Strategy

The GCSS has a cloud-resolving model intercomparison component. Modeling workshops have been conducted by the Working Group on Boundary Layer Clouds. Non-precipitating stratocumulus clouds in idealized environments were examined using Large Eddy Simulation models (Moeng et al. 1995).

The GCSS Working Group on Precipitating Convective Cloud Systems has an ongoing model intercomparison based on convection over the tropical western Pacific. The data set used in the model evaluation is from the Tropical Ocean Global Atmosphere Coupled Ocean Atmosphere Response Experiment (TOGA COARE). To identify scientific and numerical issues as well as to minimize the complications and difficulties of modeling precipitating cloud systems, prototype numerical experiments were conducted (e.g. Grabowski et al. 1996a). This working group intends to move on to continental cloud systems in due course. The GCIP ESOP in 1996, that focused on the GCIP Large Scale Area-South West (LSA-SW) during the warm season, is an opportunity to study organized precipitating systems. A prototype experiment relating to GCIP could start as soon as adequate resources are available and the ESOP data have been analyzed.

GCSS/GCIP Projects

The following are candidate projects. Additional projects may arise; for example, noting that the 1997 GCIP ESOP will concentrate on wintertime processes, a GCSS initiative on frontal clouds is a possibility (Ron Stewart, private communication).

Project 1: Investigate the coupling of surface and boundary layer processes with convection under the influence of evolving large-scale forcing.

Comprehensive modeling studies of convection over the tropical oceans have been performed. Grabowski et al. (1996a) and Xu and Randall (1996) demonstrated, in simulations of convection during the GARP Atlantic Tropical Experiment (GATE), that realistic life cycles and transports could be achieved using two-dimensional cloud resolving models. This has been extended to three dimensions by Grabowski et al. A 39-day simulation of TOGA COARE convection (Wu et al. 1996) is equally encouraging.

Since the convective life cycle over land is quite different from that over the ocean, 2D modeling should be undertaken over the GCIP region (e.g. a domain of ~900km in the horizontal by ~40km in the vertical) to examine the coupling of convection with the boundary layer and surface processes--- that is to add a precipitating cloud component to existing GCIP studies. A key issue will be the treatment in these coarse-grid models of the atmospheric boundary layer in convectively-disturbed conditions. This could involve two GCSS Working Groups (Boundary Layer and Precipitating Convective Cloud Systems). The precipitating convection study could progress to three-dimensional simulations (e.g. domain of ~400km in the horizontal by ~400km by ~40km in the vertical).

Project 2: Quantify uncertainties in NWP models associated with precipitating convective cloud systems.

An issue to be explored is the large-scale effect of organized cloud systems, which are ubiquitous over the U.S. Southern Great Plains. These systems are copious (but intermittent) producers of precipitation over a large-area because of their longevity and propagation. Consequently, they have a significant hydrologic impact; they affect the surface fluxes; and they are likely to be responsive to changes in the large-scale circulation (e.g., through the influence on convection of vertical shear which may change in response to variability, on various time scales, in the low-level nocturnal jet originating from the Gulf of Mexico).

These organized systems violate the scale-separation assumption underpinning present parameterization methods. Organized fluxes are not adequately treated in existing convective parameterization schemes. For example, it has been shown that large mesoscale systems in the tropical western Pacific cause uncertainties in a medium-range NWP model (Moncrieff and Klinker 1997), mainly because the part-resolution causes an over-prediction of the thermodynamic and momentum tendencies.

Project 3: Quantify the large-scale effects of organized convection

Cloud-resolving models have been successfully employed to determine the transport properties by organized convection in idealized tropical western Pacific environments (Wu and Moncrieff 1996). A modeling and analysis study over the continental U.S., recognizing the very different role of the boundary layer over continental land masses from over the ocean, would be a valuable addition to existing knowledge. Interactively-nested, three-dimensional models (e.g. Clark and Farley 1984), containing microphysical and surface flux parameterizations would be used to simulate organized convection over the GCIP/ARM domain.

The CSU Regional Area Modeling System (RAMS) is another interactively-nested model being used to devise parameterizations of mesoscale convective systems (MCSs). The mesoscale parameterization is tied to a version of the Arakawa-Schubert convective parameterization scheme which is modified to employ a prognostic closure. One of the two MCS case studies being used is from the central U.S. ( Alexander and Cotton 1995).

Moncrieff (1992) addressed the poorly-understood issue of convective momentum transport at a basic level by formulating a dynamical model of the mass and momentum fluxes, and also pointed the way to its parameterization in large-scale models. LeMone and Moncrieff (1994) evaluated the fluxes predicted from this model against observations. Liu and Moncrieff (1996) added the effects of shear and buoyancy to the archetypal model. As far as GCIP is concerned, a possible course of action is to evaluate how well these dynamical models represent the mass and momentum fluxes by squall line convection over the Southern Great Plains. This could be a stand-alone project but, preferably, should be conducted as part of the analysis of cloud-resolving model data sets.

The GCIP science plan (WMO 1992) identified one particular field campaign that cut across several GCIP scientific objectives. The year long field effort (with embedded IOPs) would be used to validate the largescale application of surface-atmosphere flux calculation models forced by remote sensing data, standard meteorological observations, and analyses thereof. This project would provide the following missing components, which are directly relevant to the large-scale objectives of GCIP:

• Time-series fields of evaporation, with a spatial-resolution on the order of a few kilometers and temporal resolution of hours to days.

• Time-series fields of the surface radiation budget (same spatial-temporal resolution as above).

• Time-series fields of soil moisture, with a spatial resolution of a few kilometers and a temporal resolution of days to weeks.

The provision of these additional quantities would not only close the water and energy budget equations for the region but would also provide more detailed information on the spatial distributions of moisture and energy sinks and sources within the experimental area. Measurement and modeling techniques developed with ISLSCP over the last five years could be used to address these missing components.

NOAA has already started a contribution to this effort with a new flux tower operating since May, 1995 in the Little Washita area of Oklahoma. Also augmentation of a flux tower at Oak Ridge, Tennessee has occurred and a third flux tower was added in 1996 at Bondville, Illinois.

In keeping with the philosophy of an effective, directed but economic field effort the following measurements are proposed.

(i) Four to six flux towers should be located within the GCIP area. These will be sited on the basis of a land cover/climatological classification of the GCIP area, conducted well ahead of time, using AVHRR data among other sources. The flux towers should be located near the (monitoring) radiation rigs and should measure:

Latent heat flux
Sensible heat flux
Shear stress
Soil heat flux

These measurements should be made throughout one experiment year, preferably several years.

(ii) Airborne eddy correlation

Eddy correlation aircraft (preferably twin engine aircraft like the NCAR King Air or the NAS/NRC Twin Otter) will be used during a series of Intensive Field Campaigns (IFC); perhaps three or four IFC's each of 10-20 days during the experimental year.

The aircraft will be used to conduct the following tasks:

• Measurement of fluxes over 30x30 km areas of homogeneous surface conditions centered on the flux sites.

• Measurement of fluxes over long low-level transects across gradients of soil moisture/vegetation conditions; preferably between flux sites and in conjunction with Landsat/SPOT/AVHRR acquisitions.

• Measurement of divergence/gradient terms using 'box pattern' flight lines centered on the flux sites.

These airborne eddy correlation data will be used to validate the large-scale application of surface-atmosphere flux calculation models forced by remote sensing data and meteorological observations or analyses.

(iii) Airborne soil moisture measurements

Aircraft equipped with gamma-ray or microwave sensors should be used to make soil moisture transect measurements. In some cases, these should be validated by a compact ground measurement exercise.

The routinely-acquired satellite data and the combined surface observations/analysis fields of meteorological conditions will be used to drive regional scale models that will calculate continuous time-series fields of the following quantities:

Radiation:

Insolation, PAR
Absorbed insolation, Absorbed PAR, Albedo
Downward longwave
Emitted longwave
Net radiation


Heat Fluxes:

Latent heat flux (evapotranspiration)
Sensible heat flux
Ground heat flux

Momentum:

Shear stress (roughness length)

Surface conditions:

Soil moisture
Vegetation state (FPAR)

Given the fact that the ARM program is investigating radiative transfer processes in the atmosphere as its highest priority at a site within the GCIP study area, GCIP will continue to collaborate with ARM via the existing ARM/GCIP/ISLSCP working group. However, there is a need for GCIP to take a more active role in developing a new joint focus of interest between ARM and GCIP in the area of measuring and modeling the warm season convective production of clouds and precipitation. This is an emerging joint interest of high priority to both scientific programs that should be addressed as a collaborative initiative over the next few years.

8.3 Collaboration with NASA Initiatives in the Mississippi River Basin

Several aspects of the NASA program relate direct to priority science of GCIP. The field studies on soil moisture in the ARM/CART region in 1997 relate directly to some of the science discussed in Section 6, and active collaboration should be sought between GCIP coupled modeling scientists and NASA observational scientists to secure maximum scientific benefit from that study. Equally, NASA and NOAA share an interest in providing improved management of water resources in the GCIP LSA-E, most probably through the Tennessee Valley Authority. Both agencies also share an important common interest in documenting, understanding and, to the extent possible, predicting seasonal-to-interannual variability in the southwest monsoon season, and evaluating the consequences of that variability on the vulnerable human management systems in that region.

8.4 Collaboration with PACS and GOALS

GCIP is working with the Pan-American Climate Studies (PACS) portion of the GOALS Program to develop a plan for joint studies centered on the North American monsoon system. Such research will include interfacing regional coupled atmosphere-land system models with global coupled ocean-atmosphere models as an important scientific focus.

8.5 Collaboration with the US Weather Research Program

The US Weather Research Program (USWRP), which is jointly funded by NOAA and NSF, has as one of its major goals the development of techniques to improve quantitative precipitation forecasts over short time scales. As part of this process the USWRP has been holding small workshops on relevant issues including precipitation prediction. GCIP is exploring areas of common interest to the USWRP with a view to initiating some joint studies in precipitation estimation and prediction. The data collection for ESOP-95 was carried out as a joint undertaking with the USWRP WAVE project.

8.6 FAA Collaboration - The Water Vapor Sensing System (WVSS) for Commercial Aircraft

The commercial aircraft high resolution sounding will provide winds, temperature, and water vapor (discussed below). Such profiles will aid the research goals stated in Section 5 concerning the ability to improve water balance calculations with soundings at a far greater frequency than twice per day. Such water vapor profiles will also contribute to the precipitation research discussed in Section 6.

The development of a water vapor sensing system (WVSS) for commercial air carriers was funded by the FAA under the Commercial Aviation Sensing Humidity (CASH) Program. NOAA's Office of Global Programs (GCIP) is now co-funding the procurement phase with the FAA (which is now called the WVSS program). A competitive contract was awarded to Lockheed Martin Corporation (LMC) in July 1995.

1998 Activities

Data from the first commercial aircraft with the WVSS was continuously available for the last four months of WY97. We are achieving an excellent dynamic range of mixing ratio information, the data comparison between ascent and descent from the same air terminal has been consistent, and comparisons against radiosondes (when possible) have been quite good with consistent vertical changes in degrees of wetness. All of our error quality control checks also appear to be working as we gave them to Allied Signal.

This first aircraft was a United Parcel Service (UPS) B-757. FAA certification for all B-757 was finally granted in August 1997. Installation for the next five WVSS units on UPS aircraft began in September 1997 and is expected to be completed in November 1997. The government exercised an option in the LMC contract for 60 additional units in September 1997.

A radiosonde intercomparison test of the six WVSS units and radiosondes launched at Louisville, KY (there is currently no radiosonde site at the UPS hub), will occur in November 1997. This is being co-funded by this project and the National Weather Service. The UPS labor strike, other FAA delays, and the reduced GCIP budget for this program have all contributed to a delay in the government's decision on the next 60 units. Subsequently, the production team for the WVSS must be reassembled and the delivery of the next 60 units can only begin in February 1998 and be completed by the end of fiscal year 1998. However, considerable work can be accomplished with the first six WVSS units (they provide 24 ascents and 24 descents per day) toward data evaluation and new 4DDA algorithms for water vapor information being combined with satellite data. By the end of the year the 66 WVSS units will be providing 264 new water vapor profiles each day, primarily over the GCIP region. The descent and en route data will also contribute to the satellite calibration and water vapor flux calculations.

Evaluation of the data will be performed by NOAA Forecast Systems Laboratory (FSL) for GCIP and the FAA. Quality-controlled data sets of wind, temperature, and water vapor from the commercial aircraft will be made available through the GCIP in situ data source module described in Section 9.

1999 and 2000 Activities

A second contract option for an additional 100 WVSS units will be made in FY98 as originally planned; however, due to GCIP budget decreases for this program, the full amount of money for this option is not available. The current price for this option is approximately $9K per unit and $6K for installation. The FAA has agreed to continue their share in FY99 and thus pay

back a lender of funds in FY98. If GCIP continues to fund this program at the current rate (including a similar FY99 commitment), then approximately 80 of the 100 units can be procured. This would provide the last two years of GCIP with water vapor information from 146 aircraft. Moreover, the WVSS can be continued into the future for a continued GCIP/PACS project as once the capital costs for the WVSS have been paid, the operational and maintenance cost are trivial.

8.7 Cooperative Atmospheric-Surface Exchange Study (CASES)