Assessing Phase 2 NMME (NMME-2) forecasts for improved seasonal predictions of drought and water management

“The over-arching goal of the proposed project is to assess and document the NMME Phase-2 seasonal forecasts for hydrological seasonal forecasts of drought and water management and compare these to the skill from the Phase-1. The availability under NMME-2 of a broader suite of variables, and daily forecasts rather than monthly, offers the potential for significant improvements that need to be documented through forecast experiments and systematic analysis.

Over the last 15 years NOAA’s Climate Program Office (CPO) has fostered the development of objective hydrologic and drought monitoring through support for the North American Land Data Assimilation System (NLDAS), experimental drought and hydrological forecasting using CFSv1 and more recently CFSv2, and the experimental North American Multi-Model Ensemble project (NMME). PI Wood has participated in these developments through supported projects in NLDAS, the NCEP’s Climate Test Bed and NMME. PI Wood’s current drought and hydrological seasonal forecast system is based on NMME Phase 1 (NMME-1), which utilizes monthly mean forecasts as inputs to his VIC land surface model and produces drought indices (SPI, a soil moisture based drought index, river discharge and seasonal temperature index).

As NMME transitions into phase 2 (NMME-2), there are two major differences that will impact its use for hydrological seasonal forecasting and related applications in drought and water management: the higher temporal resolution of the available forecasts – primarily daily – and equally important the availability of many more forecast variables, including surface meteorology and energy variables that can be used to force hydrological models, and surface water budget variables such as soil moisture, runoff and evapotranspiration.

Proposed Research. The higher temporal resolution of the forecasts and the new forecast variables in the NMME-2 archive offers a number of opportunities for advancing seasonal drought and hydrological forecasting. These include the following:
1. Given that NMME-2 provides high temporal resolution forecast products for many additional variables, how will these new products contribute to improved drought and hydrologic seasonal forecasts?
2. To what extent will seasonal forecast skill from NMME-2 predictions change when compared to predictions based on NMME-1? For the first time ever, the variables needed as inputs to hydrological models will be offered at high temporal resolutions from seasonal forecast models. This allows the impact on drought and water availability forecasts to be assessed and quantified.
3. How can drought and hydrologic forecasts for month 1 be improved using NMME-2? The PI’s group has tested the usefulness of the first 14 day forecast from GEFSv2 in a seasonal hydrologic forecast system based on monthly CFSv2 forecasts. During the proposed project, similar experiments with NMME-2 will be carried out within a multi-model framework, either with or without GEFSv2 for weeks 1-2. This will advance the usefulness of a seamless forecast system.

Relevance to the MAPP Program. Assessing whether NMME-2 will lead to improved seasonal forecast applications and skill is central to determining its usefulness in NCEP operations – especially for the needs of NIDIS and the Climate Prediction Center. Additionally, the needs of the private sector for seasonal forecasts are growing, and assessing the usefulness of NMME-2 for water management and agricultural crop management as well as extremes (drought and wet periods; heat waves and cold seasons) is critical. The proposed project will offer research results that can address these needs.”