) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites..
water droplets (Image credit: Brianna Santellan on Unsplash)
Precipitation Prediction Grand Challenge Strategy
Introduction to the PPGC Strategy
Strategy objectives, actions, and impacts
When will it rain? Where will it rain? How much will it rain? Will the precipitation fall as rain, snow, or sleet? The answers to these questions are needed by every person and business in the United States and at almost every timescale, spanning from the next hour to daily to decadal.
Many extreme events and the impacts they cause (e.g., billion-dollar disasters) are associated with precipitation amounts (either too much or too little). A key to reducing these impacts is to be able to anticipate when, where and how much precipitation will fall. Although models have a good track record of simulating global and regional temperature, precipitation related fields are not captured as well. The challenges in forecasting precipitation and the ongoing user needs have been acknowledged by the U.S. Congress and the Executive Office of the President with several established mandates, including the 2017 Weather Act, and the administration’s 2021 Earth System Predictability priority led by Office of Science and Technology Policy (OSTP). NOAA has recently launched the Precipitation Prediction Grand Challenge (PPGC) Initiative to help further align NOAA’s research efforts in the coming years.
Strategic Goal
Provide more accurate, reliable, and timely precipitation forecasts across timescales from weather to subseasonal-to-seasonal (S2S) to seasonal-to-decadal (S2D) through the development and application of a fully coupled Earth system prediction model.
In order to make significant progress on the Precipitation Prediction Grand Challenge, the strategy must address seven critical questions:
- What are the major systematic errors in precipitation prediction systems and the ultimate causes?
- Which key physical processes most strongly impact model biases and precipitation prediction?
- How can systematic errors in precipitation forecasts be prioritized and fixed?
- What lessons can be learned from previous successes as well as previous mistakes?
- What major knowledge gaps persist regarding precipitation processes and prediction?
- What new capabilities and opportunities could reduce major systematic errors in and improve precipitation prediction?
- How can NOAA best organize activities across the federal, academic, and private research enterprise, both nationally and internationally, to make significant progress?
Impact
NOAA’s research is aligned across all readiness levels (research, to operations, to services) and timescales, resulting in:
- improved process understanding;
- a significant reduction in systematic errors in NOAA global models;
- improved operational precipitation prediction skill; and
- effective engagement, development, and delivery of decision-support tools and services based on this improved skill.
To learn more about the six strategic objectives and supporting actions over the next five years, we encourage you to read through the full strategy.
By leveraging major progress from the last several decades, and by learning lessons from the past, the Working Group recommends six ambitious strategic objectives to provide more accurate, reliable, and timely precipitation forecasts across timescales through development and application of NOAA’s fully coupled Earth System prediction model.
Objective 1. Enhance and sustain user engagement
Acknowledging the necessary emphasis on science to improve prediction, a deep and continuous connection to the end user is key to building products and services that are useful, usable, and actually used (see also NWS Strategic Plan, Goal 1, sections 1.1-1.6; NESDIS Reimagined User Engagement Pillar 4 of 5; and NOAA Model for Service Delivery).
Action 1.1 Strengthen existing user engagement entities to continuously engage with internal and external end users and partners to understand user needs, translate their requirements, co-develop products, deliver and train them. Users include forecasters, researchers, observing teams, interagency, international, and also sector-specific, internal and external users.
Action 1.2 Champion co-production of precipitation applications between user engagement entities and product development teams to ensure a relevant and usable response to internal and external user needs. Also, examine how existing products can be used or modified.
Action 1.3 Develop and sustain region-specific networks to understand users and their needs and build the capability and capacity to respond to the needs across timescales to inform place-based, sector-based decisions.
Impacts
NOAA has prioritized and strengthened the connection between the user and the development of information, products, and services that are useful, usable, and actually used. NOAA has formalized a framework outlining the specific use of user engagement as central to product and services development and delivery.
Objective 2. Improve precipitation prediction products and applications
Action 2.1 Revamp precipitation products and services to effectively communicate uncertainty, with particular focus on translating model forecasts into services that are tuned to the impacts decision makers care about.
Action 2.2 Post process raw coupled model precipitation forecasts to calibrate for bias and other deficiencies and to quantify prediction uncertainties.
Action 2.3 Translate precipitation output into actionable visualizations and data that help forecasters more directly make a forecast decision.
Action 2.4 Establish reforecasts and high quality multi-decadal analyses of precipitation to support more statistically advanced precipitation post-processing techniques.
Action 2.5 Design prediction verification metrics based on physical reasoning and user applications.
Impacts
NOAA will have a coordinated post-processing effort, with all collaborating groups using common observation and analysis datasets, and community-developed post-processing and verification software tools for calibration and verification. Improvements in precipitation forecasts will be realized by ‘front line’ decision makers (e.g., improved public health and safety by improving emergency planning and actions; improved national security by better assessing likely outcomes) served by NOAA and our partners.
Objective 3. Improve prediction system for precipitation
These improvements will address large-scale systematic errors in NOAA’s Unified Forecast System and guide development and improvement of precipitation prediction capabilities across time-scales.
Action 3.1 Improve Unified Forecast System (UFS) Precipitation Forecasts by addressing errors from initialization. This must go hand in hand with investigation of particular locations and variables in initial conditions that precipitation prediction is most sensitive to. Precipitation challenges due to UFS initialization include:
- Understand and quantify error growth in UFS models and its attribution to the inaccuracy and gaps of initial conditions;
- Use existing observing technologies and innovative new technologies to fill critical observational gaps;
- Develop targeted coupled (e.g. ocean-land-sea-atmosphere) and atmospheric (e.g., radar, all-sky, lightning, etc) data assimilation capabilities that take full advantage of existing and forthcoming observations;
- Develop reanalysis and reforecast capabilities to ensure consistency between initial conditions and models required to maximize hindcast skill and calibrate model output.
Action 3.2 Improve Unified Forecast System (UFS) Precipitation Forecasts by addressing errors from model biases. This must go hand in hand with investigation of particular physical processes critical to precipitation prediction (Action 3.3). Similar to many other models, the UFS precipitation biases include:
- Precipitation distribution, with underestimation of heavy precipitation and overestimation of very light precipitation;
- Diurnal cycle of precipitation, with maxima too early in the day;
- Initiation of convective precipitation, for example due to errors in boundary layer stability;
- Double ITCZ;
- Organized tropical convective precipitation features (such as the Madden–Julian oscillation (MJO)) are poorly represented.
Bias correction will benefit from optimizing the prediction system design, including use of large ensembles, optimal resolution, efficient computing, and innovative coupling/nesting/adaptive configurations.
Action 3.3 Improve physics in coupled air-sea-land-ice models by emphasizing co-development of all model components, focusing on UFS. This requires the following:
- Promote more effective use of process-level diagnostics by operational model developers;
- Identify opportunities to remove key deficiencies in UFS models through integrated modeling-observational studies of coupled processes and model studies focusing especially on global, regional, and local hydrological cycles;.
- Implement innovative physics packages (e.g., scale-aware convective parameterizations, sophisticated microphysics and boundary layer schemes) that directly lead to reduction in systematic errors in precipitation prediction;
- Explore new technologies for modeling, including artificial intelligence, machine learning, stochastic perturbation, global and cloud resolving models.
- Identify and experiment in a number of targeted domains to accelerate research and development.
Impacts
NOAA will at least double the rate of improvement of precipitation prediction skill in operations, improve model competencies especially for extremes of drought or flood, and improve ensemble systems to better characterize uncertainty in its forecasts.
Objective 4. Sustain, enhance, and exploit observations
Ensure that all available high-quality observations are utilized and critical observational gaps are identified and bridged. A continuum of activities using observations is required, including initializing and constraining models, deriving data products, and focusing research on testing models and improving prediction capabilities. Continuous, high-quality, scientifically sound, global-scale observations of a number of environmental variables are critical. The Precipitation Prediction Grand Challenge will identify the observations required to improve precipitation prediction and also optimize the use of existing observing systems, recognizing resource, technological, and logistical limitations.
Action 4.1 Extend observations needed to advance understanding of precipitation predictability across space and time scales, and for variables including soil moisture, sea surface temperature/salinity/density (sea-surface properties), vertical profiles of upper-ocean heat and freshwater contents, ocean turbulence, sea ice, snow cover, vegetation, air-sea fluxes, OCTOBER 30, 2020 | PAGE 19 OF 44 as well as the air surface and air vertical profiles of buoyancy, stability, precipitation, cloud properties, temperature, wind, and moisture content (liquid, ice, vapor). Predictions of precipitation beyond a few weeks also depend on the state of long-term oscillations in the 3-D atmosphere-oceanland system, which involve changes in precipitation, circulation, and air-sealand-sea ice interactions.
Action 4.2 Enhance observations needed for advancing understanding of physical processes key to precipitation prediction by exploiting new technologies (such as new satellite sensors and observation strategies, and in situ platforms) and new product development to support prediction model development and improvement.
Action 4.3 Sustain, enhance, and exploit observations for improving initial conditions for precipitation prediction by maintaining, enhancing, and innovating existing and future observational capabilities (including space and ground remote sensing, aircraft, ships, autonomous devices, coastlines, offshore platforms, moorings), and by supporting international observing efforts. The number of surface ocean, surface land, and ocean-profiling observation platforms in operation today does not cover all areas or time windows (i.e. domain and frequency of observations) needed for initializing or validating precipitation prediction models. The current network of observations requires maintenance and also support for expansion and innovations of technology in regions determined to be critical for precipitation predictability. While new remote sensing methods for observing the upper atmosphere must continue to be explored, the current atmospheric sounding observation system also needs reinvestment to meet its critical original goals. An assessment of an optimal space and surface observing architecture for the purposes of this initiative is recommended. This should also consider forthcoming geostationary hyperspectral assets which will provide more frequent and spatially continuous soundings.
Action 4.4 Expand observations needed for calibration, verification, and uncertainty quantification of precipitation prediction products and ensure that these effectively align with the requirements and needs across NOAA and with our partners. An example is satellite radiances, which require continuous calibration and validation against in situ observations, direct interpretation in geophysical terms, and assimilation into reanalysis models. NOAA should ensure that precipitation prediction skill and applications are optimized because of sufficient use of observations, which should include their transformation into Climate Data Records to ensure their long term homogeneity.
Impacts
NOAA ensures that all available high quality observations are utilized in the end-to-end precipitation prediction forecasts and their verification. NOAA tests and employs new observation capabilities in a timely manner within NOAA testbeds, including emerging technologies, to help fill gaps in observations critical to precipitation prediction.
Objective 5. Improve process-level understanding and modeling
These improvements will provide insights into the key processes that have the strongest imprint on model biases and the limits of precipitation predictability and prediction skill.
Action 5.1 Support synthesizing analysis of existing observations (in situ, remote, routine data, UxS and citizen science observations, and field experiments) to ensure that researchers and modelers effectively extract new knowledge from existing datasets to advance process understanding and modeling.
Action 5.2 Identify and fix key model deficiencies and processes that contribute the most to error growth leading to systematic biases and limited precipitation prediction skill, with particular emphasis on reformulating convective parameterizations to work across space and timescales. Other examples include large-scale dynamics versus local-scale processes, coupling of atmosphere, ocean, sea ice and land, physical processes of clouds and boundary layers, and errors in initial conditions.
Action 5.3 Conduct targeted field experiments to obtain intensive observations to better understand the key processes important to precipitation in partnership with the United States and international research communities. While a unified/integrated approach is recommended, specific objectives ought to be targeted at particular areas and lead times with broad impacts.
Action 5.4 Conduct a hierarchy of modeling-observational integrated studies spanning from large-eddy simulations, cloud-resolving models, super-parameterization approaches, and global models having a range of resolutions.
Action 5.5 Focus studies on extremes (flooding and drought on land, as well as precipitation anomalies and events over ocean) via synergistic observational, modeling studies to better understand the physical bounds on precipitation extremes, and how addressing model and observational deficiencies can most effectively improve prediction and preparedness.
Impacts
NOAA has identified and is addressing key model biases that limit precipitation forecasts. NOAA is using process-level model-observation integrated approaches to improve model representation of key processes for prediction (e.g., the diurnal cycle of convection in the tropics) and the leading patterns of variability (including blocking patterns, teleconnections, and the Madden Julian Oscillation) that influence model characterization of precipitation, especially for intense rainfall or sudden-onset drought/flood.
Objective 6. Advance understanding of precipitation predictability
These advances will address what is practically predictable and the theoretical limits of predictability. Addressing what is practicably predictable requires understanding the needs of decision makers and incorporating that understanding into Earth system predictability research.
Action 6.1 Modernize observational and modeling tools for the study of predictability. Identify, improve, and focus coordinated modeling and observations on key sources of error. Explore how observations can be included in predictability studies, and what would be the best way to conduct predictability studies using models with biases. New tools can include AI (artificial intelligence) and ML (machine learning), stochastic perturbations to the entire physical space, global cloud or cloud-resolving models permitting.
Action 6.2 Understand precipitation predictability, its sources and barriers, including dependence on time, locations and scales, and windows of opportunity (temporal high predictability). The focus will be on the sources of predictability that have the strongest influences on precipitation prediction, especially extremes.
Action 6.3 Expand the definition of predictability to be directly applicable to users. For example, include rainfall and other relevant fields critical to applications for the regions and sectors.
Impacts
- NOAA and the broader community has a better understanding of what is practically predictable and NOAA is using that to guide planning of future investments in process studies and process modeling (Objective 4). NOAA can also set realistic targets for the expectations of researchers, product developers and end users.
- NOAA is applying improved understanding of predictability to improve our interpretation of model errors, such as ENSO false alarm rate, errors in steering flow, or regime transitions.
- NOAA has an improved understanding of the needs of decision makers and is incorporating that understanding into Earth system predictability research. Improved understanding of predictability will also improve our communication of uncertainty to stakeholders.
- NOAA has an improved understanding of the limits of precipitation predictability and is using that to reduce errors in initial conditions.
Below is the vision for progress toward the objectives of this plan [master link to objectives section] at the milestones of two, five, and ten years.
In Two Years
{Objective 3} NOAA has developed coherent, testable hypotheses for major precipitation systematic errors in partnership with the United States and international communities, and has funded research projects to test these hypotheses and to explore productive algorithmic changes to address them.
{Objective 1} NOAA is supporting partnerships with the decision-making community to establish precipitation predictability metrics that are of relevance to users and applications.
{Objective 4} NOAA has supported projects that synthesize existing field observations for more effective applications to prediction model development and improvement.
{Objective 5} NOAA is working with the broad national and international community to implement multi-disciplinary, multi-agency process studies targeting key deficiencies in forecast systems that limit precipitation prediction skill the most.
{Objective 4} NOAA is working with the broad national and international community to identify gaps in the global observing system that contribute the most to errors in initial conditions of forecast systems which limit precipitation prediction skill.
{Objective 3} NOAA will have implemented a global reanalysis/reforecast strategy to support the post processing of convection permitting forecasts using reforecasts and high quality multi-decadal analyses.
{Objective 4} NOAA will have assessed the feasibility of generating high-resolution precipitation analyses over many years with its Multi-Resolution / Multi-Sensor (MRMS) system and will be prepared for product generation, in support of precipitation post processing, situational awareness, forecast validation, and more.
{Objective 3} NOAA will have GEFSv12 reforecast-based algorithmic modifications delivered to the Meteorological Development Laboratory to provide improved precipitation forecasts in its National Blend of Models program.
{Objective 3} NOAA supports community-based UFS coupled global modeling and will apply the Global Energy and Water Exchanges (GEWEX) coupling metrics to investigate how coupling affects precipitation prediction and related diagnostics.
{Objective 3} NOAA supports development of the UFS regional convection-allowing-ensemble forecast system, capable of improved accuracy and forecast spread.
{Objective 3} NOAA has invested its research resources in model-observation integrated studies to advance understanding and model improvement.
{Objective 2} NOAA has established more robust communities of practice and communication between managers, researchers, forecasters, and stakeholders to ensure that the most pressing needs for improved precipitation forecast information are met.
{Objective 1} New investments are driven by stakeholder needs for information and applications.
{implicit in the others} NOAA has completed 25% of the actions proposed in this Strategy.
In Five Years
{implicit in the others} NOAA has successfully allocated substantial resources to support activities of the Precipitation Prediction Grand Challenge.
{Objective 1} NOAA has further strengthened robust communities of practice and hosted multiple engagements amongst managers, researchers, forecasters, and the broader community to meet the needs for improved precipitation forecast information.
{Objective 1} NOAA is leveraging robust communities of practice to guide the production and application of precipitation products and research results.
{Objective 2} NOAA has improved products (visualization and product design) to address forecaster’s and partners needs, including information on confidence and alternative scenarios.
{Objective 3} NOAA’s UFS data assimilation is able to fully capture the previously untapped information provided from cloudy/precipitating radiances.
{Objective 4} NOAA will have extended its MRMS high-resolution precipitation analyses back as far as 2013 in support of precipitation post processing.
{Objective 3} NOAA will have implemented algorithms for use of the GEFSv12 data in its National Blend of Models precipitation post-processing algorithm.
{Objective 3} NOAA has reduced major systematic biases in UFS models by margins that are statistically significant.
{Objective 4} NOAA has implemented new observational networks using advanced observing technologies in critical areas that have the highest impact on reductions in initial condition errors for precipitation prediction.
{Objective 6} NOAA’s new investments in research, observations, modeling and research to operations have led to clear advances in understanding of precipitation predictability, improvements in precipitation prediction skill in operations, and effective dissemination of prediction products to their users.
{Objective 2} NOAA “portals” provide easy access to precipitation forecasts, products, and applications across timescales.
{Objective 3} NOAA will have applied global ensemble reforecasts and high quality multi-decadal analyses of precipitation in producing regional to local probabilistic precipitation forecasts to inform decision making and risk management across time scales from days to 1 month.
{Objective 4} NOAA is leveraging past and future investments to make the underlying observations available.
{Objective 3} NOAA has increased the rate of improvement of the skill of precipitation forecasts from 15% per decade to 30% per decade (i.e., doubled the rate of improvement per decade).
{Objective 3} NOAA will have implemented improved convective parameterizations in its global models that provide more realistic propagation of coherent precipitation features such as the Madden-Julian Oscillation and mesoscale convective systems.
{Objective 3} NOAA will have implemented improvements in its global and regional systems that result in improved representation of the diurnal cycle of convection, developing and dissipating in the forecast model at the same time as in precipitation analyses.
{implicit in the others} NOAA has accomplished 50% of the actions proposed in this Strategy.
In Ten Years
{ALL} NOAA has further increased the rate of improvement of the skill of precipitation forecasts from 30% per decade to 50% per decade (i.e., double the skill over two decades).
{ALL} NOAA provides water resource managers, emergency management, community leaders, managers, researchers, forecasters, and the public improved robust early warnings and longer-term predictions for issues ranging from life-saving actions, such as evacuations and swift-water rescues caused by extreme precipitation events, to water resource management affected floods and droughts and supports broader decision-making applications.
{implicit in the others} NOAA has completed 80% of the actions proposed in this Strategy.
{ALL} NOAA’s precipitation prediction is viewed as the best worldwide.
The PPGC Working Group will build the partnerships both nationally and internationally across disciplines, agencies, and organizations that are needed to achieve the goals, actions and objectives recommended in this strategy.
- Continue to work through the NOAA Weather, Water, and Climate Board to complete the strategy and move towards implementation.
- Adopt the Weather, Water, and Climate Board’s Service Delivery Framework (Sept 2020) and use it to frame the PPGC actions in the future.
- Coordinate and collaborate with interagency partners (e.g., NASA, DOE, NSF, USGS, DoD, USDA, USBR) through USGCRP.
- Respecting the role and processes of NOAA’s Science Advisory Board, collaborate with appropriate groups to identify opportunities for coordination in identifying broad user needs and support for decision making applications.
- Leverage the Executive Office of the President’s interagency Council for Advancing Meteorological Services for coordination.
- Engage with the World Meteorological Organization (WMO), including:
- World Weather Research Program;
- World Climate Research Program;
- Global Climate Observing System;
- Coordination Group for Meteorological Satellites’ International Precipitation Working Group and other relevant Coordination Group for Meteorological Satellites Science Working groups;
- Committee on Earth Observation Satellite’s Precipitation Virtual Constellation;
- Research Board on Weather, Climate, Water and the Environment; and
- Scientific Advisory Panel.
The Precipitation Prediction Grand Challenge Strategy is a living document that will be updated on a regular basis based on new inputs.
In November 2019, the Weather, Water, and Climate Board convened the cross-Line Office Precipitation Prediction Grand Challenge Working Group and charged the members with developing NOAA’s strategy for accelerating improvements to global models that will lead to increases in precipitation prediction skill and that will translate these improvements into more effective decision support.
The Working Group began meeting in December 2019, with an initial mandate to write the detailed outline. Based on that outline, six writing teams were identified. The writing teams reported on progress and sticking points at the bi-monthly Working Group meetings that were held from January through July 2020 and each writing team developed a paper on their topic:
-
User Needsoffsite link offsite link. Identifying the tools, data, and information required by decision makers under changing environmental conditions.
-
Model Limitationsoffsite link offsite link. Identifying the current model limitations and prioritizing the systematic errors in the models that should be addressed first.
-
Research Questionsoffsite link offsite link. Identifying and prioritizing the focused research questions that need to be addressed.
-
Transitionsoffsite link offsite link. Articulating how to accelerate R2O2R [research-to-operations-to-research].
-
Observationsoffsite link offsite link. Identifying current observation systems that could be better leveraged and those that need to be implemented.
-
Post Processingoffsite link offsite link. Identifying post-processing approaches to ensure that model output is transformed into actionable decision support services.
When will it rain? Where will it rain? How much will it rain? Will the precipitation fall as rain, snow, or sleet? The answers to these questions are needed by every person and business in the United States and at almost every timescale, spanning from the next hour to daily to decadal.
Many extreme events and the impacts they cause (e.g., billion-dollar disasters) are associated with precipitation amounts (either too much or too little). A key to reducing these impacts is to be able to anticipate when, where and how much precipitation will fall. Although models have a good track record of simulating global and regional temperature, precipitation related fields are not captured as well. The challenges in forecasting precipitation and the ongoing user needs have been acknowledged by the U.S. Congress and the Executive Office of the President with several established mandates, including the 2017 Weather Act, and the administration’s 2021 Earth System Predictability priority led by Office of Science and Technology Policy (OSTP). NOAA has recently launched the Precipitation Prediction Grand Challenge (PPGC) Initiative to help further align NOAA’s research efforts in the coming years.
Strategic Goal
Provide more accurate, reliable, and timely precipitation forecasts across timescales from weather to subseasonal-to-seasonal (S2S) to seasonal-to-decadal (S2D) through the development and application of a fully coupled Earth system prediction model.
In order to make significant progress on the Precipitation Prediction Grand Challenge, the strategy must address seven critical questions:
- What are the major systematic errors in precipitation prediction systems and the ultimate causes?
- Which key physical processes most strongly impact model biases and precipitation prediction?
- How can systematic errors in precipitation forecasts be prioritized and fixed?
- What lessons can be learned from previous successes as well as previous mistakes?
- What major knowledge gaps persist regarding precipitation processes and prediction?
- What new capabilities and opportunities could reduce major systematic errors in and improve precipitation prediction?
- How can NOAA best organize activities across the federal, academic, and private research enterprise, both nationally and internationally, to make significant progress?
Impact
NOAA’s research is aligned across all readiness levels (research, to operations, to services) and timescales, resulting in:
- improved process understanding;
- a significant reduction in systematic errors in NOAA global models;
- improved operational precipitation prediction skill; and
- effective engagement, development, and delivery of decision-support tools and services based on this improved skill.
To learn more about the six strategic objectives and supporting actions over the next five years, we encourage you to read through the full strategy.
By leveraging major progress from the last several decades, and by learning lessons from the past, the Working Group recommends six ambitious strategic objectives to provide more accurate, reliable, and timely precipitation forecasts across timescales through development and application of NOAA’s fully coupled Earth System prediction model.
Objective 1. Enhance and sustain user engagement
Acknowledging the necessary emphasis on science to improve prediction, a deep and continuous connection to the end user is key to building products and services that are useful, usable, and actually used (see also NWS Strategic Plan, Goal 1, sections 1.1-1.6; NESDIS Reimagined User Engagement Pillar 4 of 5; and NOAA Model for Service Delivery).
Action 1.1 Strengthen existing user engagement entities to continuously engage with internal and external end users and partners to understand user needs, translate their requirements, co-develop products, deliver and train them. Users include forecasters, researchers, observing teams, interagency, international, and also sector-specific, internal and external users.
Action 1.2 Champion co-production of precipitation applications between user engagement entities and product development teams to ensure a relevant and usable response to internal and external user needs. Also, examine how existing products can be used or modified.
Action 1.3 Develop and sustain region-specific networks to understand users and their needs and build the capability and capacity to respond to the needs across timescales to inform place-based, sector-based decisions.
Impacts
NOAA has prioritized and strengthened the connection between the user and the development of information, products, and services that are useful, usable, and actually used. NOAA has formalized a framework outlining the specific use of user engagement as central to product and services development and delivery.
Objective 2. Improve precipitation prediction products and applications
Action 2.1 Revamp precipitation products and services to effectively communicate uncertainty, with particular focus on translating model forecasts into services that are tuned to the impacts decision makers care about.
Action 2.2 Post process raw coupled model precipitation forecasts to calibrate for bias and other deficiencies and to quantify prediction uncertainties.
Action 2.3 Translate precipitation output into actionable visualizations and data that help forecasters more directly make a forecast decision.
Action 2.4 Establish reforecasts and high quality multi-decadal analyses of precipitation to support more statistically advanced precipitation post-processing techniques.
Action 2.5 Design prediction verification metrics based on physical reasoning and user applications.
Impacts
NOAA will have a coordinated post-processing effort, with all collaborating groups using common observation and analysis datasets, and community-developed post-processing and verification software tools for calibration and verification. Improvements in precipitation forecasts will be realized by ‘front line’ decision makers (e.g., improved public health and safety by improving emergency planning and actions; improved national security by better assessing likely outcomes) served by NOAA and our partners.
Objective 3. Improve prediction system for precipitation
These improvements will address large-scale systematic errors in NOAA’s Unified Forecast System and guide development and improvement of precipitation prediction capabilities across time-scales.
Action 3.1 Improve Unified Forecast System (UFS) Precipitation Forecasts by addressing errors from initialization. This must go hand in hand with investigation of particular locations and variables in initial conditions that precipitation prediction is most sensitive to. Precipitation challenges due to UFS initialization include:
- Understand and quantify error growth in UFS models and its attribution to the inaccuracy and gaps of initial conditions;
- Use existing observing technologies and innovative new technologies to fill critical observational gaps;
- Develop targeted coupled (e.g. ocean-land-sea-atmosphere) and atmospheric (e.g., radar, all-sky, lightning, etc) data assimilation capabilities that take full advantage of existing and forthcoming observations;
- Develop reanalysis and reforecast capabilities to ensure consistency between initial conditions and models required to maximize hindcast skill and calibrate model output.
Action 3.2 Improve Unified Forecast System (UFS) Precipitation Forecasts by addressing errors from model biases. This must go hand in hand with investigation of particular physical processes critical to precipitation prediction (Action 3.3). Similar to many other models, the UFS precipitation biases include:
- Precipitation distribution, with underestimation of heavy precipitation and overestimation of very light precipitation;
- Diurnal cycle of precipitation, with maxima too early in the day;
- Initiation of convective precipitation, for example due to errors in boundary layer stability;
- Double ITCZ;
- Organized tropical convective precipitation features (such as the Madden–Julian oscillation (MJO)) are poorly represented.
Bias correction will benefit from optimizing the prediction system design, including use of large ensembles, optimal resolution, efficient computing, and innovative coupling/nesting/adaptive configurations.
Action 3.3 Improve physics in coupled air-sea-land-ice models by emphasizing co-development of all model components, focusing on UFS. This requires the following:
- Promote more effective use of process-level diagnostics by operational model developers;
- Identify opportunities to remove key deficiencies in UFS models through integrated modeling-observational studies of coupled processes and model studies focusing especially on global, regional, and local hydrological cycles;.
- Implement innovative physics packages (e.g., scale-aware convective parameterizations, sophisticated microphysics and boundary layer schemes) that directly lead to reduction in systematic errors in precipitation prediction;
- Explore new technologies for modeling, including artificial intelligence, machine learning, stochastic perturbation, global and cloud resolving models.
- Identify and experiment in a number of targeted domains to accelerate research and development.
Impacts
NOAA will at least double the rate of improvement of precipitation prediction skill in operations, improve model competencies especially for extremes of drought or flood, and improve ensemble systems to better characterize uncertainty in its forecasts.
Objective 4. Sustain, enhance, and exploit observations
Ensure that all available high-quality observations are utilized and critical observational gaps are identified and bridged. A continuum of activities using observations is required, including initializing and constraining models, deriving data products, and focusing research on testing models and improving prediction capabilities. Continuous, high-quality, scientifically sound, global-scale observations of a number of environmental variables are critical. The Precipitation Prediction Grand Challenge will identify the observations required to improve precipitation prediction and also optimize the use of existing observing systems, recognizing resource, technological, and logistical limitations.
Action 4.1 Extend observations needed to advance understanding of precipitation predictability across space and time scales, and for variables including soil moisture, sea surface temperature/salinity/density (sea-surface properties), vertical profiles of upper-ocean heat and freshwater contents, ocean turbulence, sea ice, snow cover, vegetation, air-sea fluxes, OCTOBER 30, 2020 | PAGE 19 OF 44 as well as the air surface and air vertical profiles of buoyancy, stability, precipitation, cloud properties, temperature, wind, and moisture content (liquid, ice, vapor). Predictions of precipitation beyond a few weeks also depend on the state of long-term oscillations in the 3-D atmosphere-oceanland system, which involve changes in precipitation, circulation, and air-sealand-sea ice interactions.
Action 4.2 Enhance observations needed for advancing understanding of physical processes key to precipitation prediction by exploiting new technologies (such as new satellite sensors and observation strategies, and in situ platforms) and new product development to support prediction model development and improvement.
Action 4.3 Sustain, enhance, and exploit observations for improving initial conditions for precipitation prediction by maintaining, enhancing, and innovating existing and future observational capabilities (including space and ground remote sensing, aircraft, ships, autonomous devices, coastlines, offshore platforms, moorings), and by supporting international observing efforts. The number of surface ocean, surface land, and ocean-profiling observation platforms in operation today does not cover all areas or time windows (i.e. domain and frequency of observations) needed for initializing or validating precipitation prediction models. The current network of observations requires maintenance and also support for expansion and innovations of technology in regions determined to be critical for precipitation predictability. While new remote sensing methods for observing the upper atmosphere must continue to be explored, the current atmospheric sounding observation system also needs reinvestment to meet its critical original goals. An assessment of an optimal space and surface observing architecture for the purposes of this initiative is recommended. This should also consider forthcoming geostationary hyperspectral assets which will provide more frequent and spatially continuous soundings.
Action 4.4 Expand observations needed for calibration, verification, and uncertainty quantification of precipitation prediction products and ensure that these effectively align with the requirements and needs across NOAA and with our partners. An example is satellite radiances, which require continuous calibration and validation against in situ observations, direct interpretation in geophysical terms, and assimilation into reanalysis models. NOAA should ensure that precipitation prediction skill and applications are optimized because of sufficient use of observations, which should include their transformation into Climate Data Records to ensure their long term homogeneity.
Impacts
NOAA ensures that all available high quality observations are utilized in the end-to-end precipitation prediction forecasts and their verification. NOAA tests and employs new observation capabilities in a timely manner within NOAA testbeds, including emerging technologies, to help fill gaps in observations critical to precipitation prediction.
Objective 5. Improve process-level understanding and modeling
These improvements will provide insights into the key processes that have the strongest imprint on model biases and the limits of precipitation predictability and prediction skill.
Action 5.1 Support synthesizing analysis of existing observations (in situ, remote, routine data, UxS and citizen science observations, and field experiments) to ensure that researchers and modelers effectively extract new knowledge from existing datasets to advance process understanding and modeling.
Action 5.2 Identify and fix key model deficiencies and processes that contribute the most to error growth leading to systematic biases and limited precipitation prediction skill, with particular emphasis on reformulating convective parameterizations to work across space and timescales. Other examples include large-scale dynamics versus local-scale processes, coupling of atmosphere, ocean, sea ice and land, physical processes of clouds and boundary layers, and errors in initial conditions.
Action 5.3 Conduct targeted field experiments to obtain intensive observations to better understand the key processes important to precipitation in partnership with the United States and international research communities. While a unified/integrated approach is recommended, specific objectives ought to be targeted at particular areas and lead times with broad impacts.
Action 5.4 Conduct a hierarchy of modeling-observational integrated studies spanning from large-eddy simulations, cloud-resolving models, super-parameterization approaches, and global models having a range of resolutions.
Action 5.5 Focus studies on extremes (flooding and drought on land, as well as precipitation anomalies and events over ocean) via synergistic observational, modeling studies to better understand the physical bounds on precipitation extremes, and how addressing model and observational deficiencies can most effectively improve prediction and preparedness.
Impacts
NOAA has identified and is addressing key model biases that limit precipitation forecasts. NOAA is using process-level model-observation integrated approaches to improve model representation of key processes for prediction (e.g., the diurnal cycle of convection in the tropics) and the leading patterns of variability (including blocking patterns, teleconnections, and the Madden Julian Oscillation) that influence model characterization of precipitation, especially for intense rainfall or sudden-onset drought/flood.
Objective 6. Advance understanding of precipitation predictability
These advances will address what is practically predictable and the theoretical limits of predictability. Addressing what is practicably predictable requires understanding the needs of decision makers and incorporating that understanding into Earth system predictability research.
Action 6.1 Modernize observational and modeling tools for the study of predictability. Identify, improve, and focus coordinated modeling and observations on key sources of error. Explore how observations can be included in predictability studies, and what would be the best way to conduct predictability studies using models with biases. New tools can include AI (artificial intelligence) and ML (machine learning), stochastic perturbations to the entire physical space, global cloud or cloud-resolving models permitting.
Action 6.2 Understand precipitation predictability, its sources and barriers, including dependence on time, locations and scales, and windows of opportunity (temporal high predictability). The focus will be on the sources of predictability that have the strongest influences on precipitation prediction, especially extremes.
Action 6.3 Expand the definition of predictability to be directly applicable to users. For example, include rainfall and other relevant fields critical to applications for the regions and sectors.
Impacts
- NOAA and the broader community has a better understanding of what is practically predictable and NOAA is using that to guide planning of future investments in process studies and process modeling (Objective 4). NOAA can also set realistic targets for the expectations of researchers, product developers and end users.
- NOAA is applying improved understanding of predictability to improve our interpretation of model errors, such as ENSO false alarm rate, errors in steering flow, or regime transitions.
- NOAA has an improved understanding of the needs of decision makers and is incorporating that understanding into Earth system predictability research. Improved understanding of predictability will also improve our communication of uncertainty to stakeholders.
- NOAA has an improved understanding of the limits of precipitation predictability and is using that to reduce errors in initial conditions.
Below is the vision for progress toward the objectives of this plan [master link to objectives section] at the milestones of two, five, and ten years.
In Two Years
{Objective 3} NOAA has developed coherent, testable hypotheses for major precipitation systematic errors in partnership with the United States and international communities, and has funded research projects to test these hypotheses and to explore productive algorithmic changes to address them.
{Objective 1} NOAA is supporting partnerships with the decision-making community to establish precipitation predictability metrics that are of relevance to users and applications.
{Objective 4} NOAA has supported projects that synthesize existing field observations for more effective applications to prediction model development and improvement.
{Objective 5} NOAA is working with the broad national and international community to implement multi-disciplinary, multi-agency process studies targeting key deficiencies in forecast systems that limit precipitation prediction skill the most.
{Objective 4} NOAA is working with the broad national and international community to identify gaps in the global observing system that contribute the most to errors in initial conditions of forecast systems which limit precipitation prediction skill.
{Objective 3} NOAA will have implemented a global reanalysis/reforecast strategy to support the post processing of convection permitting forecasts using reforecasts and high quality multi-decadal analyses.
{Objective 4} NOAA will have assessed the feasibility of generating high-resolution precipitation analyses over many years with its Multi-Resolution / Multi-Sensor (MRMS) system and will be prepared for product generation, in support of precipitation post processing, situational awareness, forecast validation, and more.
{Objective 3} NOAA will have GEFSv12 reforecast-based algorithmic modifications delivered to the Meteorological Development Laboratory to provide improved precipitation forecasts in its National Blend of Models program.
{Objective 3} NOAA supports community-based UFS coupled global modeling and will apply the Global Energy and Water Exchanges (GEWEX) coupling metrics to investigate how coupling affects precipitation prediction and related diagnostics.
{Objective 3} NOAA supports development of the UFS regional convection-allowing-ensemble forecast system, capable of improved accuracy and forecast spread.
{Objective 3} NOAA has invested its research resources in model-observation integrated studies to advance understanding and model improvement.
{Objective 2} NOAA has established more robust communities of practice and communication between managers, researchers, forecasters, and stakeholders to ensure that the most pressing needs for improved precipitation forecast information are met.
{Objective 1} New investments are driven by stakeholder needs for information and applications.
{implicit in the others} NOAA has completed 25% of the actions proposed in this Strategy.
In Five Years
{implicit in the others} NOAA has successfully allocated substantial resources to support activities of the Precipitation Prediction Grand Challenge.
{Objective 1} NOAA has further strengthened robust communities of practice and hosted multiple engagements amongst managers, researchers, forecasters, and the broader community to meet the needs for improved precipitation forecast information.
{Objective 1} NOAA is leveraging robust communities of practice to guide the production and application of precipitation products and research results.
{Objective 2} NOAA has improved products (visualization and product design) to address forecaster’s and partners needs, including information on confidence and alternative scenarios.
{Objective 3} NOAA’s UFS data assimilation is able to fully capture the previously untapped information provided from cloudy/precipitating radiances.
{Objective 4} NOAA will have extended its MRMS high-resolution precipitation analyses back as far as 2013 in support of precipitation post processing.
{Objective 3} NOAA will have implemented algorithms for use of the GEFSv12 data in its National Blend of Models precipitation post-processing algorithm.
{Objective 3} NOAA has reduced major systematic biases in UFS models by margins that are statistically significant.
{Objective 4} NOAA has implemented new observational networks using advanced observing technologies in critical areas that have the highest impact on reductions in initial condition errors for precipitation prediction.
{Objective 6} NOAA’s new investments in research, observations, modeling and research to operations have led to clear advances in understanding of precipitation predictability, improvements in precipitation prediction skill in operations, and effective dissemination of prediction products to their users.
{Objective 2} NOAA “portals” provide easy access to precipitation forecasts, products, and applications across timescales.
{Objective 3} NOAA will have applied global ensemble reforecasts and high quality multi-decadal analyses of precipitation in producing regional to local probabilistic precipitation forecasts to inform decision making and risk management across time scales from days to 1 month.
{Objective 4} NOAA is leveraging past and future investments to make the underlying observations available.
{Objective 3} NOAA has increased the rate of improvement of the skill of precipitation forecasts from 15% per decade to 30% per decade (i.e., doubled the rate of improvement per decade).
{Objective 3} NOAA will have implemented improved convective parameterizations in its global models that provide more realistic propagation of coherent precipitation features such as the Madden-Julian Oscillation and mesoscale convective systems.
{Objective 3} NOAA will have implemented improvements in its global and regional systems that result in improved representation of the diurnal cycle of convection, developing and dissipating in the forecast model at the same time as in precipitation analyses.
{implicit in the others} NOAA has accomplished 50% of the actions proposed in this Strategy.
In Ten Years
{ALL} NOAA has further increased the rate of improvement of the skill of precipitation forecasts from 30% per decade to 50% per decade (i.e., double the skill over two decades).
{ALL} NOAA provides water resource managers, emergency management, community leaders, managers, researchers, forecasters, and the public improved robust early warnings and longer-term predictions for issues ranging from life-saving actions, such as evacuations and swift-water rescues caused by extreme precipitation events, to water resource management affected floods and droughts and supports broader decision-making applications.
{implicit in the others} NOAA has completed 80% of the actions proposed in this Strategy.
{ALL} NOAA’s precipitation prediction is viewed as the best worldwide.
The PPGC Working Group will build the partnerships both nationally and internationally across disciplines, agencies, and organizations that are needed to achieve the goals, actions and objectives recommended in this strategy.
- Continue to work through the NOAA Weather, Water, and Climate Board to complete the strategy and move towards implementation.
- Adopt the Weather, Water, and Climate Board’s Service Delivery Framework (Sept 2020) and use it to frame the PPGC actions in the future.
- Coordinate and collaborate with interagency partners (e.g., NASA, DOE, NSF, USGS, DoD, USDA, USBR) through USGCRP.
- Respecting the role and processes of NOAA’s Science Advisory Board, collaborate with appropriate groups to identify opportunities for coordination in identifying broad user needs and support for decision making applications.
- Leverage the Executive Office of the President’s interagency Council for Advancing Meteorological Services for coordination.
- Engage with the World Meteorological Organization (WMO), including:
- World Weather Research Program;
- World Climate Research Program;
- Global Climate Observing System;
- Coordination Group for Meteorological Satellites’ International Precipitation Working Group and other relevant Coordination Group for Meteorological Satellites Science Working groups;
- Committee on Earth Observation Satellite’s Precipitation Virtual Constellation;
- Research Board on Weather, Climate, Water and the Environment; and
- Scientific Advisory Panel.
The Precipitation Prediction Grand Challenge Strategy is a living document that will be updated on a regular basis based on new inputs.
In November 2019, the Weather, Water, and Climate Board convened the cross-Line Office Precipitation Prediction Grand Challenge Working Group and charged the members with developing NOAA’s strategy for accelerating improvements to global models that will lead to increases in precipitation prediction skill and that will translate these improvements into more effective decision support.
The Working Group began meeting in December 2019, with an initial mandate to write the detailed outline. Based on that outline, six writing teams were identified. The writing teams reported on progress and sticking points at the bi-monthly Working Group meetings that were held from January through July 2020 and each writing team developed a paper on their topic:
-
User Needsoffsite link offsite link. Identifying the tools, data, and information required by decision makers under changing environmental conditions.
-
Model Limitationsoffsite link offsite link. Identifying the current model limitations and prioritizing the systematic errors in the models that should be addressed first.
-
Research Questionsoffsite link offsite link. Identifying and prioritizing the focused research questions that need to be addressed.
-
Transitionsoffsite link offsite link. Articulating how to accelerate R2O2R [research-to-operations-to-research].
-
Observationsoffsite link offsite link. Identifying current observation systems that could be better leveraged and those that need to be implemented.
-
Post Processingoffsite link offsite link. Identifying post-processing approaches to ensure that model output is transformed into actionable decision support services.