NOAA Water Resources Prediction

1

The Interstate Council on Water Policy/National Water Supply Alliance Washington, DC Roundtable Meeting

April 3, 2019

Presentation Outline

•NWS Strategic Vision and Plan

•National Water Center

•Key Partnerships

•National Water Model

•Hydrologic Ensemble Forecast Service

•Summary

2

TO

O M

UC

HP

OO

R Q

UA

LIT

YT

OO

LIT

TLE

3Involves the entire US Weather, Water and Climate Enterprise working together

NOAA NWS Strategic Vision/Societal Outcome: A Weather-Ready Nation

• Society is prepared for and responds to weather, water, and climate dependent events

• Build community resiliency in the face of increasing vulnerability to extreme weather, water, and climate events

4

NOAA NWS Strategic Plan 2019-2022: Building a Weather-Ready Nation

Released March 26, 2019

https://www.weather.gov/media/wrn/NWS_Weather-Ready-Nation_Strategic_Plan_2019-2022.pdf

Overarching Goals:• Reduce the impacts of weather, water, and

climate events by transforming the way people receive, understand, and act on information

• Harness cutting-edge science, technology, and engineering to provide the best observations, forecasts, and warnings

• Evolve the NWS to excel in the face of change through investment in our people, partnerships, and organizational performance

5

NOAA NWS Strategic Plan 2019-2022: Water-Specific Goals

https://www.weather.gov/media/wrn/NWS_Weather-Ready-Nation_Strategic_Plan_2019-2022.pdf

• Deliver actionable water resources information from national to street-level and across all time scales;

• Provide minutes-to-months river forecasts that quantify both atmospheric and hydrologic uncertainty;

• Improve forecasts of total water in the coastal zone by linking terrestrial and coastal models in partnership with the National Ocean Service; and

• Deliver forecasts of flood inundation linked with other geospatial information to inform life-saving decisions.

National Water Center

6

• Center of excellence for water resources science and prediction

• Catalyst to transform water prediction through enterprise collaboration

• Operations Center for water resources common operating picture and decision support services on all time scales

Initial Operating Capacity: May 26, 2015

NWC has hosted more than 80 scientific meetings with over 2800 participants

March 21, 2019

• Federal Agencies including Integrated Water Resources Science & Services (IWRSS)

Water Information and Science

US Army Corpsof Engineers

Water Prediction

7

Water Management

FEMA Response and Mitigation

Key Partnerships

• Academia/Research including National Science Foundation, CUAHSI, UCAR, NCAR

• Water Resources Managers, Emergency Managers, and other Enterprise Stakeholders

Setting the Stage for Transformation

Centralized Water Forecasting Demo

(2015)

National Water Model (NWM) Development and Demonstration

Centralized Water Resources Data Services

Water Resources Test and Evaluation Service

Enhanced Water Prediction Capability

(2016)

Hyper-Resolution Modeling

Real-Time Flood Forecast Inundation Mapping

Enhance Impact-Based Water Resources Decision Support Services

Integrated Water Prediction

(2017)

Stand up the NWC Operations Center

Increase HPC capacity

Couple terrestrial freshwater and coastal estuary models for total water prediction

National Water ModelV1.0 Implemented August 16, 2016

•Continental-scale water resources model providing high resolution, spatially continuous estimates of major water cycle components

•Operational forecast streamflow guidance for currently underserved locations: 3,600 forecast points 2.7 million stream reaches (>700 fold increase in spatial density)

Current NWS River Forecast Pointsoverlaid with NWM Stream Reaches

National Water Model V1.2Analysis and Forecast Cycling Configurations

Cycling Forecast Forcing Outputs

All configurations include reservoirs (1260 water bodies parameterized with level pool scheme)

Analysis assimilates >7,000 USGS Observations10

Analysis & Short-Range

Medium-Range

Long-Range

Hourly

4 x Day

Daily Ensemble

(16 members)

18 hours

10 days

30 days

MRMS QPEDownscaled HRRR/RAP

Blend

Downscaled GFS

Downscaled and Bias-Corrected

CFS

1km Land States,250m Sfc Routed

Water,NHDPlus

Streamflow

1km Land States,250m Sfc Routed

Water,NHDPlus

Streamflow

1km Land States,NHDPlus

Streamflow

Foundation Established August 2016

Water Resource Model for 2.7 Million Stream Reaches

First/Second Upgrade May 2017/March 2018

Increased cycling freq. and forecast length, improved calibration,

soil/snow physics and stream DA

Third UpgradeMay 2019

Expansion to Hawaii, medium range ensembles, enhanced snow

physics, compound channel parameterization, increased

modularity, improved calibration, longer Analysis w/MPE

Fourth UpgradeFall 2020

Expansion to Puerto Rico and Great Lakes, increased modularity, enhanced reservoir

module, physics improvements, forcing bias-correction, improved calibration, and

improved Hawaii QPE11

Future Plans: Upgrading to NWM V2.0 and Beyond

Hurricane Florence

Iowa Flooding Texas Flooding

NWM Medium-Range Forecast (06Z 9/18)Lynches River at Effingham, SC

NWM Medium-Range Forecast (00Z 9/22)Iowa River at Wapello, IA

NWM Medium-Range Forecast (12Z 10/12)Trinity River at Dallas, TX

NWM Medium-Range Forecast (00Z 10/21)Nueces River near Three Rivers, TX

NWM Medium-Range Forecast (00Z 9/21)East Fork Des Moines River, Near Algona IA

NWM Medium-Range Forecast (06Z 9/15)Cape Fear River at William O. Huske Lock near Tar Heel, NC

NWM V2.0 displayed good performance for Hurricane Florence flooding, and in Iowa and Texas flood events,new ensemble begins to capture forecast uncertainty

NWM V2.0 Medium-Range Real-time Ensemble Forecast Examples

21

NOAA-USGS National Water Model CollaborationSupported by the USGS Water Prediction Work Program

• Observations: Gap analysis for next generation observing systemin support of water prediction (e.g., streamflow, stream temperature, soil moisture, etc.)

• Establishment of Community Modeling Development Environment

• Development and application of hydroinformatics for integration of geospatial data and of decision support tools

• Co-development of enhanced water prediction capabilities (on NHD+ channel network)

² Streamflow

² Temperature (Water Quality)

² Sediment/Constituent Entrainment -- sources, characteristics, and movement of materials (Water Quality)

² Constituent Transport -- including physical and chemical fate (Water Quality)

• Maps supported emergency management efforts to stage supplies in non-flooded areas and to target relief efforts

• Texas Division of Emergency Management needed information on exisiting and maximum possible flood extent

Experimental National Water Model GuidanceHurricane Harvey: Flood Inundation Mapping

Houston

Port Arthur

Beaumont

15

Experimental National Water Model GuidanceHurricane Harvey: Forecast Time to Bankfull Flow

18 Hour Forecast Valid 18Z August 27, 2017

Houston

Buffalo Bayou

16

Experimental National Water Model GuidanceHurricane Harvey: Forecast Time to Bankfull Flow

18 Hour Forecast Valid 18Z August 27, 2017

Houston

Buffalo Bayou

PEAK INUNDATION EXTENTR&R and NWM: 5-Day Forecast

Reference Time:August 26th, 2017 at 12:00 UTC

RFC Replace and Route FIMNational Water Model FIM

Experimental National Water Model GuidanceHurricane Harvey: Flood Inundation Mapping

Houston

PEAK INUNDATION EXTENTR&R and NWM: 5-Day Forecast

Reference Time:August 26th, 2017 at 12:00 UTC

RFC Replace and Route FIMNational Water Model FIM

Experimental National Water Model GuidanceHurricane Harvey: Flood Inundation Mapping

Houston

NWM Streamflow Forecast for Hurricane Harvey10 Day Forecast (left) and corresponding Analysis (right): 12Z Aug 22 – 12Z Sep 1

Overall extreme streamflow pattern forecast several days in advance by NWM

NWM Streamflow Forecast with Global Forecast System (GFS) Precipitation Overlay

NWM Streamflow Analysis with Multi-Radar Multi-Sensor (MRMS) Precipitation Overlay

19

20

Water Resources Services - Existing and NewCurrent: Ensemble Streamflow Prediction (ESP)

● Historical climate forcings● No accounting for hydrologic

uncertainty or bias ● Suitable for long-range

forecasting only● limited number of graphical

products

Near-term: Hydrologic Ensemble Forecasting Service (HEFS)

● Utilizes short, medium, and long range weather forecasts

● Captures total uncertainty and corrects for biases in forcingsand streamflow

● Suitable for forecasting hours to years

● Wide array of data and user products

Future Services: NWM long-range ensemble prediction

● Potential to provide probabilistic forecasts at 2.7 million stream reaches

Advanced Hydrologic Prediction Service

2water.weather.gov

• 1091 river locations display HEFS experimental short and/or long-range products

• Over next few years, goal is to expand the number of HEFS river service locations to 2350

• Near tern enhancements include beginning the use of GEFSv12 model forcings in 2020

Hydrologic Ensemble Forecast Service (HEFS)

The forecast for 2pm on 12 March, HEFS shows ~5% chance the river level will reach/exceed 11 ft (well above Major Flooding level)

The forecast for 2pm on 12 March, HEFS shows ~25% chance of reaching or exceeding 7.3 ft (Moderate Flooding)

The forecast for 2pm on 12 March, HEFS shows ~75% chance of reaching or exceeding 6.5 ft (Minor Flooding)

Summary

22

• NOAA’s Water Services are Evolving– Continental scale modeling approach producing consistent, “street-level,” summit-to-

sea information to address growing stakeholder needs – Deliver comprehensive, integrated actionable water predictions/intelligence– Enhanced, extended-range ensemble predictions which quantify forecast uncertainty

• Implementing State-of-the-Art Technical Approach– Water resources prediction through state-of-the-science earth system modeling in a

high performance computing environment– Impact-based decision support services underpinned by geo-intelligence

• New Organization, Cornerstone Facility and Philosophy– Office of Water Prediction/National Water Center– Collaboration across NOAA and with Federal Partners, Academia, and the broader

Water Resources Enterprise is critical to success

Backup Slides

National Water ModelInitial Operating Capability:Model Chain

5. Channel & ReservoirRoutingModules

4. NHDPlus CatchmentAggregation(2.7M uniquecatchments andriver reaches)

2. NoahMP LSM (1 km grid)

1. NWM Forcing Engine(1 km grid)

3. Terrain RoutingModule(250 m grid)

2-way coupling

ForecastProducts

NWM uses NCAR supported community WRF-Hydro systemNWM: http://water.noaa.gov/about/nwm

WRF-Hydro: https://www.ral.ucar.edu/projects/wrf_hydro

• Updates– Addition of Hawaii to NWM domain (including Analysis and 60-hr Short-Range forecast)– Addition of Extended Analysis configuration (daily 28-hour lookback)– Addition of separate Long-Range Analysis configuration to initialize Long-Range forecasts– Addition of Medium Range ensemble forecast configuration

• 7 members 4x day (mem1=current GFS to 10 days, mem2-7=time lagged GFS to 8.5 days)– Use of 13km GFS forcing (versus 0.25 degree)– Improved downscaling of GFS and CFS forcing via Mountain Mapper approach– Improved physics (out-of-bank flow via compound channel, improved snow physics)– Improved waterbody parameters from 30m DEM, inclusion of 3,995 additional reservoirs– Improved calibration of hydrologic parameters, expanded from 1,100 to 1,400 basins– Corrections to stream connectivity– Improved code modularity– Bugzilla fixes and file metadata updates

NWM Version 2.0 Enhancements

25

New Hawaii Domain

National Water Center Annual Innovators ProgramPartnership with the academic community via Interagency Agreement with the NSF and CUAHSI to host a competitive Summer Institute

•Year one included 44 graduate students from 19 Universities, June - July 2015 – Demonstrated ability to simultaneously model the entire continental United States river network

at high spatial resolution, in near real-time for 2.7 million stream reaches

•Year two included 34 graduate students from 21 Universities, June - July 2016 – Demonstrated the ability to generate flood inundation maps utilizing NWM output – Engaged social scientists and stakeholders from the Fire, Police and Emergency Management

Communities to explore ways to best communicate water information

•Year three includes 34 graduate students from 25 Universities, June - July 2017 – Refine the recently developed process to create flood inundation maps nationally in real time – Develop a strategy for a hyper-resolution nest of the NWM – Improve the communication of water resources information

Summer Institute 201823 graduate students from 18 Universities

• Explore shallow aquifer/groundwater coupling in the National Water Model ü Improving upon NWM forecasts by investigating and quantifying groundwater-surface water

exchanges.

• Conduct hyper-resolution modelingü Improving the accuracy and analyzing the sensitivity of flood inundation predictions from hyper-

resolution models in complex urban watersheds.

• Investigate computational aspects of NWM & citizen science dataü Evaluating the significance of channel geometry and routing schemes in the context of the NWM, and

methods for integrating crowd-sourced data for improved stage and flow predictions.

FY18-19 DOC/NOAA Agency Priority Goal

28

Goal Statement: By September 30, 2019, NOAA National Weather Service will improve its flood related decision support services by (1) demonstrating a new flood inundation mapping capability serving 25 million people (approximately 8%* of the U.S. continental population) residing in flood-vulnerable freshwater basins, and (2) delivering an enhanced excessive rainfall outlook product, with lead time of “High Risk” predictions extended from two days to three days. Emergency Managers will use this information to more effectively mitigate flood impacts by prepositioning resources, ensuring critical infrastructure (e.g., hospitals, evacuation routes, etc.) are viable, and ordering evacuations.

*Future out-year goal is to incrementally expand flood inundation mapping to near 100% of the continental U.S. population residing in flood-vulnerable freshwater basins.

Mitigate Flood Impacts by Demonstrating Improved Decision Support Services to Emergency Managers

Community Advisory Committee for theOffice of Water Prediction (CAC-WP)

29

Purpose:• Conduct a thorough independent review of OWP’s water modeling capabilities with

emphasis on the National Water Model (NWM), other modeling innovations, and related data and information services.

• Bring independent expertise and perspectives from across the community to provide recommendations to improve OWP’s water modeling capabilities and related data and information services.

• Consider the various activities OWP has already undertaken to address documented requirements and associated science and service gaps.

Composition and Scope:• Two-Co Chairs and 12+ member committee comprised primarily of hydrologists, civil

engineers, and other water resources science and data science experts.

• Administratively managed by UCAR

• Any review and recommendations will be those of committee members and not of UCAR

• Meet every 12-18 months and produce a written report of its findings and recommendations.

Community Advisory Committee for theOffice of Water Prediction (CAC-WP)

30

The CAC-WP scope is broad and initial review provided advice regarding: 1. The National Water Model, including:

a. The current and future elements of the NWM as described multi-year strategic science and services framework, as well as future adjustments to that framework;

b. In-situ and remotely-sensed observations for assimilation and validation;c. Physiographic data sets such as terrain data, stream network, land use land cover, soils data,

reservoir characteristics, and other relevant data sets;d. Hydrometeorological forcings; e. Improved representation of physical processes; f. Accounting for anthropogenic processes; g. The establishment of a community developmental testbed and associated governance; h. Performance metrics to assess overall model performance and objectively evaluate potential

model upgrades; i. Involvement of NWS RFCs, WFOs, and NCEP Centers in using, validating and improving the NWM; j. Requirements for HPC resources and associated implementation strategies to optimally use available

computing resources; k. Integration into a broader unified Earth System Prediction Capability (ESPC)

2. The evolution of the OWP water resources data services;

3. Integrating the broad spectrum and large volume of water resources and related geospatial information for new product development, and enhanced impact-based decision-support services.

Community Advisory Committee for theOffice of Water Prediction (CAC-WP) Membership

31

Sector Name Organization

Government

Cline, Don (Co-Chair) USGS

Farthing, Matthew USACE - ERDC

Lesmes, David DOE

Longenecker, Gene FEMA

Moglen, Glenn USDA - ARS

Nelson, Jonathon USGS

Torgersen, Tom (Ex Officio) NSF

Young, Dwane EPA

Academia

Bales, Jerad (Ex Officio) CUAHSI

Barros, Ana Duke University

Dawson, Clint UT Austin

Fan Reinfelder, Ying Rutgers University

Foufoulou Giorgio, Efi UC Irvine

Hooper, Richard Tufts University

Maidment, David (Co-Chair) UT Austin

Community Advisory Committee for theOffice of Water Prediction (CAC-WP) MembershipContinued

32

Sector Name Organization

Academia Continued

Mandli, Kyle Columbia University

Rasmussen, Roy (Ex Officio) NCAR

Tarboton, David Utah State University

NGO Mesheli, Ehab Water Institute of the Gulf

Private Sector

Ables, Matthew Kisters North America

Kopp, Steve ESRI

Nimmich, Joe Booze Allen Hamilton

Tolle, Kristen Microsoft Research

V1.0 V1.1 V1.2 V2.0

NWM v2.0 Streamflow Hourly Correlation at USGS Gauges (WY 2014-2016, NLDAS MM Forcing)

18% have cor >= 0.8 21% have cor >= 0.8 27% have cor >= 0.8 33% have cor >= 0.8

NWM v2.0 Improvement: Correlation at All USGS Gauges

● Version-over-version improvement● Pronounced at directly calibrated sites● Calibrated/validated against hourly

(previously daily) streamflow obs● Daily metrics also improve● No assimilation of USGS observations

24

Forecast Informed Reservoir Operations (FIRO) in Russian River Watershed

• Multi-Agency study on Lake Mendocino• Can we enhance reservoir operations and

use of available storage by using forecasts to inform decisions about releasing or storing water?

• HEFS forecasts are central to optimized forecast-based reservoir operations

• Supports water control manual change request for Lake Mendocino

• Process can be replicated in other watersheds

Early Applications of HEFS