A recently completed review by the NAS/NRC GEWEX Panel recommended that GCIP should focus more on the seasonal to interannual prediction problem and recommended the following scientific mission for GCIP:

"To demonstrate skill in predicting changes in water resources on time scales up to seasonal, annual, and interannual as an integral part of the climate prediction system."

The Panel further recommended some restatement of the original science objectives to more clearly focus on the seasonal to interannual prediction problem and to add an objective pertaining to data management. GCIP has adopted these modified objectives to better reflect the emphasis of the Project which has evolved over the past five years since the completion of the GCIP Science Plan (World Meteorological Organization, 1992). The GCIP objectives are:

1) Determine and explain the annual, interannual and spatial variability of the water and energy cycles within the Mississippi River basin.

2) Develop and evaluate coupled hydrologic/atmospheric models at resolutions appropriate to large-scale continental basins.

3) Develop and evaluate atmospheric, land, and coupled data assimilation schemes that incorporate both remote and in-situ observations.

4) Improve the utility of hydrologic predictions for water resources management up to seasonal and interannual time scales.

5) Provide access to comprehensive in-situ, remote sensing and model output data sets for use in GCIP research and as a benchmark for future studies.

The GCIP Implementation Plan, comprising three volumes, was completed in 1993 and 1994. Volume I of the GCIP Implementation Plan (IGPO, 1993) is the overall planning document for the Project. It addresses the organizational framework for GCIP, the observational and database needs, and the upgrades to be made to existing operational analysis and prediction streams that produce routine four-dimensional data assimilation (4DDA) analyses for the GCIP and global domains. Volume II (IGPO, 1994a) examines the elements of a GCIP research program needed to assist the research community in addressing the specific scientific questions in the GCIP Science Plan. The overall plans for data management through the duration of the GCIP Project are described in Volume III of the GCIP Implementation Plan (IGPO, 1994b).

GCIP research involves a systematic multiscale approach to accommodate physical process studies, model development, data assimilation, diagnostics, and validation topics. Such a multiscale developmental framework for the GCIP effort has three attributes:

(1) Support for a hierarchy of scales for observational work, algorithm and model development, and validation and diagnostic studies leading to a continental-scale capability.

(2) Capacity for sequential expansion to support the evolution of research themes (e.g., initial emphasis on hydrological implications of warm- season convective precipitation, moving next to issues related to midlatitude cold-season hydrology).

(3) Flexibility to develop methods and algorithms that can be applied in data-sparse areas of the globe outside the Mississippi River basin.

The understanding and modeling of a continental scale require, from the outset, consideration of nonlinear-scale interactions in the aggregation of smaller processes to the larger scale and vice versa. Progress in this area requires that methodologies be developed to represent the coupling of processes that are important in one medium (e.g., the atmosphere) to those that are important in another (e.g., the land surface). These techniques must be suitable at the resolution of operational prediction and general circulation models (GCM) (about 10 to 100 km) and hence must be capable of representing in aggregate the effects of high levels of heterogeneity in the underlying ground surface (WMO, 1992). Accordingly, the GCIP research approach addresses activities on four scales (IGPO, 1994a):

The analyses and diagnostic studies conducted on the CSA, LSA, and ISA scales will derive their data primarily from existing sources, with augmentation of some observing systems as required. A major element of the rationale for carrying out the GCIP effort in the Mississippi River basin is the potential for full utilization of a number of observing systems (e.g., wind profiles and Doppler radars) not available to the same extent anywhere else in the world. In a number of LSAs, data from the existing synoptic and climatological networks operated by the National Weather Service can be augmented by data from relatively dense climatological networks established and operated by other Federal agencies and state organizations.

To the extent possible, the SSAs will be collocated with existing research basins, for example, the Little Washita Experimental Watershed in Oklahoma operated by the U.S. Department of Agriculture. The analyses, diagnostic studies, and model development on the SSA scale will be derived from operational data sources (augmented as necessary), existing research instrument complexes, and specially designed field programs of limited duration.

A fundamental thrust of the GCIP implementation strategy is that although the developmental activities will be initiated in limited regions, they lead toward an integrated continental-scale capability. Full continental domain studies have been important in GCIP from the beginning of the EOP in 1995. Retrospective analyses and baseline studies of water and energy balance will continue to be the main focus in the near- term. In fact, as the EOP proceeds, the GCIP-derived budgets based on regional mesoscale models will likely be superior in accuracy to budget estimates from other sources. These diagnostic studies will also be valuable for validating hydrological aspects of climate model simulations and understanding planetary-scale influences on North American hydrology.

The completion of the GCIP Science Plan in early 1992 heralded the beginning of a number of major activities in GCIP that have progressed steadily over the past four years. Some of the key accomplishments during this period are summarized in the remainder of this section within the scientific/technical implementation framework as outlined in the following section.

With the interest in climate as a science over the past decade or so, computer models of the earth/atmosphere system have taken place along two separate paths. Many of the improvements in global models for weather prediction have occurred in, or in close cooperation with, the major operational analysis centers such as the U.S. National Center for Environmental Prediction (NCEP) and the European Centre for Medium Range Weather Forecasts (ECMWF). Developments in global climate models, which have their origins in the global weather models, have generally occurred in the U.S. in large research establishments such as the NOAA/Geophysical Fluid Dynamics Laboratory (GFDL), the NASA Goddard Space Flight Center (GSFC) and the National Center for Atmospheric Research (NCAR). In the early development of strategies for implementing GCIP, it was recognized that it would be necessary to draw on the strengths offered by both of these paths. A further key strategy that was adopted early in GCIP was the need to fully exploit the high resolution, limited area models that were being applied to regional weather prediction tasks through various nesting procedures in the global models.

The GCIP research activities got underway in 1993 with primary support from NOAA. The achievements to date can be grouped under four headings of data analysis, model development, diagnostics of model output, and observing system enhancements. A more comprehensive review of this activity was published in a special issue of the Journal of Geophysical Research Volume 101, Number D3, March 20, 1996.

i. Data Analysis
- Precipitation variability and extreme events: Implications for climate models and climate change
- Physically-based subgrid scale statistical parameterization of rainfall: Coupling mesoscale meteorology with small scale statistical descriptions
- Characterization and modeling of snow distribution and associated land surface hydrology over mountainous watersheds
- Snow cover as an indicator of anthropogenic change
- Analysis and modeling of the hydrological cycle using Russian data
ii. Model Development
- Dynamic land surface/atmosphere parameterization at different spatial scales for the South Platte River Drainage
- Project for Intercomparison of Land-Surface Parameterization Schemes (PILPS)
- Coupled boundary-layer formulation and land surface processes
- Using the Mesoscale Kinetic Energy (MKE) to Parameterize Subgridscale (Mesoscale) Processes in General Circulation Models
- Evaluation of GCM Land-Surface Schemes in GCIP
iii. Diagnostics of Model Output
- Summaries of North American continental-scale hydrology using ETA Model analysis/forecast products
- North American land surface/atmosphere hydrological cycle
- Prediction and diagnosis of atmospheric moisture and surface hydrology variations over North America
iv. Observing System Enhancements
- Aircraft Water Vapor Sensing System
- Automated soil moisture and temperature profile measurements at the DoE ARM/CART Southern Great Plains Site for GCIP
- Surface flux measurements in the Little Washita Watershed in Oklahoma and Bondville, Illinois
- Surface Radiation (SURFRAD) measurements at three sites in the Mississippi River basin.

1.4.3 Achievements in the Operational Centers

The GCIP Science Plan (WMO, 1992) recognized that the building of a database for GCIP scientists would be a major undertaking and that the amount and different types of data needed for GCIP studies would require an efficient data collection and management strategy.

The accomplishments to date in database development are in the areas of Pre-EOP data collection, compilation of several initial data sets, and the implementation of a distributed data management and service system. Each of these items is summarized in Part II of this Major Activities Plan.

The responsibilities of the GCIP Data Management and Service System (DMSS) are to provide data services to GCIP investigators, adapt to the evolving data requirements, and compile the information on a five-year consolidated data set at the completion of the EOP. Carrying out these responsibilities involves an implementation approach with evolutionary improvements during the different stages of GCIP.

The purpose of the Major Activities Plan is to project a description of GCIP research and associated activities over the next two to three years to preclude the need for frequent revisions to the three volumes of the GCIP Implementation Plan. The initial version of the Major Activities Plan covered the two-year period of 1995 and 1996 with an outlook for 1997 (IGPO, 1994c) and was updated last year (IGPO, 1995).

The description of planned activities is based on what should be done in an orderly progression toward the end objectives of GCIP and with a realistic assumption about the resources that will be available to do it. Adjustments are made the following year, as appropriate, to rationalize the plans with the actual resources. The adjustments are used as a starting point for projections in the following year's update.