NSSL Multiple-Radar Multiple-Sensor MRMS Decision Briefing David

NSSL Multiple-Radar / Multiple-Sensor (MRMS) Decision Briefing David Kitzmiller (OHD/HL) Greg Stumpf (OST/MDL)

Purpose of briefing Approval of the National Severe Storms Laboratory Multiple-Radar / Multiple-Sensor (MRMS) weather decision support system as an official NOAA Line Office Transition Project. 2

Users requiring advanced radar-based products 3

Background: Motivation for MRMS • Current operational warning and QPE algorithms are based on limited single-radar data Cones-of-silence, beam broadening at far ranges, terrain blockage Very limited environmental data input • MRMS provides a more accurate and complete depiction of storms, their evolution, and precipitation characteristics • Many forecasters have lost confidence in the single-radar algorithms and have resorted to time -consuming “base data analysis” Becoming unfeasible to manage the “fire hose” of information from all radars and sensors during storm outbreaks 4

Background: What is MRMS? • A scientifically-sound framework for providing more accurate, more precise, and more timely hydrometeorological decision support data versus current single-radar algorithms. Advanced techniques in data quality control, data integration, severe weather detection and diagnosis, precipitation estimation, and short-term forecasting. A 3 D/4 D high-resolution grid of radar (and other observations) and derived severe weather & precipitation estimation products. The result of 10+ years of published research, application development, and operational testing at NSSL, WFOs, RFCs, NCEP, SPC, AWC, and the NOAA Hazardous Weather Testbed (HWT) 5

Background: What is MRMS? • MRMS = Multiple-Radar / Multiple-Sensor • Multi-Radar: Exploits the overlapping coverage of the WSR-88 D network and the Level-II real-time data feeds to build a seamless rapidly-updating high-resolution three-dimensional cube of radar data. • Multi-Sensor: Objectively blends data from the multiple-radar 3 D cubes with surface, upper air, lightning, satellite, rain gauges, and NWP environmental data, to produce highly-robust decision assistance products. • Improvements demonstrated in QPE, severe weather diagnosis, warning decision efficiency, NWP, etc. 6

MRMS Outputs 3 D Reflectivity cube 3 D Azimuthal Shear cube 2 D products to support: Hydro, Aviation, NWP, Severe Weather 7

Current specs of NSSL experimental system • Ingests data from national network of WSR-88 D, Canadian, 1 tv, 2 TDWR radars • Grid resolution of 1 km x 5 minutes over the CONUS for 2 D products; and with 31 vertical levels for 3 D products • Radar data go through an automated multi-sensor quality control (QC) to censor non-weather echoes • Generating 1 -km CONUS products every 5 -min requires 48 Linux servers • Servers and drives can be quickly configured, maintained, swapped out, and extended to add capacity for Greater temporal and spatial resolution Dual Polarization data/products Additional radars (foreign, tv, mobile, PAR, CASA) 8

Current MRMS Product Distribution Operational RUC/HRRR 9

MRMS Benefits Hydrology Provides added precipitation estimation coverage in over 99% of the “radar hostile” regions (terrain, bright-band, hail). Provides measured improvements in data QC removing false precipitation echoes, which improves the RTMA, and reduces inaccurate precipitation estimates and unnecessary flash flood warnings. Greatly benefits the Western U. S. , where there is a combination of major flooding vulnerability and radar coverage gaps. “MRMS/Q 2 provides precipitation estimates from portions of southwest Texas, western New Mexico, south central Colorado and Mexico where few if any other sources of precipitation data exist. This area has now experienced major flooding in two of the past four years due to dissipating tropical systems. The loss of Q 2 would, in short, mean the loss of what we have seen to be our most accurate radar-based QPE. ” – Greg Story (West Gulf RFC). Integrates lightning sensor data to apply advanced precipitation segregation (convective, stratiform) for more accurate rainfall rates The river stage forecast errors in some basins are reduced by up to 1 meterusing MRMS/Q 2. This improved accuracy will lead to major savings in flood mitigation efforts (e. g. , sandbagging, evacuations). 10

MRMS Benefits Aviation Facilitates NOAA meeting IOC and MOC Next. Gen radar based product requirements Provides a straight forward research-tooperations (R 2 O) integration of advanced radar dependent aviation products without WSR-88 D system dependencies or delays “The MRMS system provides a flexible and efficient software computing architecture to accommodate rapid changes or additions to the Next. Gen objectives/requirements while providing a straight forward research-tooperations (RTO) integration platform for AWRP-funded, radar dependent, turbulence and icing solution portfolios without system dependencies or delays in implementation within the WSR-88 D system. . ” – FAA Reduced Weather Impact (RWI) plan A development platform already installed at FAA Tech Center will facilitate R 2 O of Next. Gen applications into the operational MRMS system. The turbulence and icing solution portfolios can be implemented and configured to allow the creators to modify and improve solutions quickly in addition to expanding the capability to utilize gap filling and international radar networks. MRMS is a viable platform to support the NOAA Convective Initiation (CI) project 11

MRMS Benefits NWP MRMS provides the only reliable, rapidly-available, quality-controlled, high-resolution 3 D radar grid covering the CONUS. The 3 D reflectivity grid is used to derive a 3 D latent heating field used to create convective cloud fields. Model forecasts are improved using the MRMS data. “… 3 D cloud and hydrometeor fields are not well sampled by conventional observing systems and no single observing platform fully captures the needed information. The NSSL MRMS data have been absolutely critical to the success we had in the radar reflectivity data assimilation in the RUC, and its impact on the HRRR, and now in the RR. ” - Steve Weygandt (GSD) Assimilation of MRMS into the RUC enables prediction of excessive rainfall amounts (up to 2” per 12 h), which are rarely generated by RUC without MRMS ingest. These major rain events would go unforecasted without MRMS data assimilation. 12

MRMS Benefits Severe Weather Improvements in warning decision efficiency (via HWT) allows more WFO resources to be directed to customer decision support during high impact events. More accurate at the depiction of hail and mesocyclones makes warnings more precise and timely. NWS Warnings HWT Warnings The hail swath and tornado track verification system provides the WFOs, Red Cross, and FEMA with rapid automated assessments of storm damage areas. “Even today, the volume of radar data alone is such that it is nearly impossible for a well-trained meteorologist to be assured they have interpreted all of the relevant information. ” - David Andra (WFO OUN SOO), commenting on the need for a robust data integration system to control the “fire hose” of multiple rapidly-updating data streams during WFO warning operations. Complete CONUS coverage can be used to develop robust climatological archives for hail, rotating storms, and precipitation, benefiting agriculture, insurers, etc. “SPC routinely uses a number of real-time national (MRMS) fields to monitor the evolution, intensity, and hazards associated with convective storms. The increased resolution of products and improved accuracy in hail size and rainfall has been noted by many forecasters here. ” - Steve Weiss (SPC SOO) 13

Short-term impacts of not transitioning the full NSSL MRMS to operations • Users must continue their dependence on a research-funded quasi-operational experimental prototype that is not maintained 24/7 and has no backup Field offices and NCEPs are increasing their reliance on MRMS products OHD 0 -6 h QPF system under development is dependent on MRMS radar input • Multiple organizations might require their own stand-up systems to support their single-agency specific operational requirements AWRP must fund the AWC Testbed to set up an interim MRMS solution now to meet FAA’s turbulence (NTDA) deliverables • FAA development system at FAA Tech Center remains without NOAAsponsored operational target • The NCEP IBM-AIX MRMS system for the RUC is limited NCEP-EMC must routinely upgrade NSSL software to IBM-AIX (requires extra staff time) Will not meet the needs of other agencies/centers (limited product suite) Product latency and i/o not adequate for short-fused warning ops 14

Long-term impacts of not transitioning the full NSSL MRMS to operations • OAR must continue to use NOAA research funding to maintain quasi-operational systems, versus advancing R&D • Major Western U. S. hydrologic limitations in “radar-hostile” regions continue • Warning decision efficiency continues to degrade and information overload grows • NWS must continue to pay for inferior (4 -bit, 2 D) commercial mosaics 15

Preferred Deployment Option NCEP/NCO Linux Cluster: Will meet multiple agency requirements Transition explored Completed draft NCEP charter for implementation Completed testing on Dell Blade server architecture and solid state drives Has proven O&M expertise and robust backup capability Eliminates NCEP staff costs for porting upgrades to IBM-AIX Latency and i/o is adequate for short-fused warning ops Will match specs of development platforms, facilitating R 2 O Cost estimate (primary & backup): One time transition: $3 million Annual O&M: $1 million 16

Alternate Deployment Options AWC/SPC FAA Tech Center Telecommunications Operations Center (TOC) Private Industry 17

In conclusion… NSSL MRMS is a sole-source, scientifically-sound solution that meets many high-priority requirements for multiple agencies, including Next. Gen. There is no other proven alternative national capability to integrate and optimize information from multiple radar networks (WSR-88 D, TDWR, ASR) into a 4 D data cube. 2 FACTOID: One new WSR-88 D costs ~$10 M and provides < 0. 13 M mi coverage An MRMS System at NCEP: * Will operate for 10 years at the same cost, a small % addition of the total cost of the NEXRAD program 2 * Will provide >3. 2 M mi total CONUS coverage at $3 / sq mi (versus $100 / sq mi for one radar) * Will increase the coverage of all WSR-88 Ds, combined, by 35% 18

Discussion Approval of the National Severe Storms Laboratory Multiple-Radar / Multiple-Sensor (MRMS) weather decision support system as an official NOAA Line Office Transition Project. 19

Backup Slides 20

What else comes with the “MRMS package”? • Development framework (like ADE, CODE) A foundation for research and product development Facilitates transition of Research to Operations • Data Options: Push Pull (subscription) Complex Data Retrievals (for locally-derived custom products) • Verification Systems Precipitation Hail and Tornado • Network monitoring system 21

MRMS Benefits NOAA and Its Users and Partners • Will provide a single-authority, operationally 24/7 maintained, seamless, scientifically-sound, high resolution 4 D data cube of integrated radar and sensor data for multiple agencies • Improvements in depictions of convective initiation, structure, and evolution for warnings, forecasts, air traffic routing • An immediate target of opportunity for Next. Gen • Will provide framework for research and development • Will provide a clear R 2 O path for transition of new science to operations • Will provide an analysis of record to more robustly understand severe weather and precipitation climatologies nationwide • Will strengthen existing and establish new partnerships with multiple development and operational agencies • Will save lives, property, aviation delays/accidents 22

Advantages of MRMS Integrate multiple-radar and multiple-sensor information No longer single-radar specific. More accuracy in detection and diagnosis (better sampling - more “eyes” looking at storms). Rapid-update capability Uses virtual volume scan concept. Better lead time (no more waiting until end of volume scan for guidance). Automatically fill in outage gaps by other sensors Provides better continuity of operations. 23

Many single radars provide many different answers VIL = 34 KJAN VIL = 52 KLIX VIL = 45 KMOB 24

Many single radars provide many different answers Is this the best detection? VIL = 52 KLIX 25

Multiple radars provide a single robust answer KJAN Or is this? KLIX KMOB 26

Single radar data Single Radar 27

Blended 3 D multi-radar data Radars in network supplement each other: Overlapping coverage Fills in gaps from terrain blockage Increased sampling frequency Multiple Radars 28

Single Radar v. Multiple Radar Vertical Cross-Section VIL Tracks 29

NMQ: National mosaic and Multi-sensor QPE http: //nmq. ou. edu • A suite of quantitative precipitation estimation algorithms • Uses multi-sensor observations • Products delivered to RFCs (AHIPS) and NWS FOs (FFMPA) for operational use • An archive a precipitation products 30

NMQ Precipitation Products • Reflectivity Products Vertical Profile of Reflectivity, Bright-band Identification Hybrid-Scan Reflectivity (VPR corrected) • Precipitation Products 1 -HR Precip, 3 -HR, 6 -HR, 12 -HR, 24 -HR, and 72 -HR Precip Radar only, Multi-Sensor, Radar with Gauge Bias Correction Local Gauge Bias Gauge-only Precip • Precipitation Verification Automated comparison to Stage-II, Stage-IV, HADS, MPE 31

NMQ Precipitation Products Improvements over current operational products: Vertical reflectivity profile correction Adapted to mountainous terrain Seamless hybrid scan minimizes effects of beam blockage on radials Superior quality control to operational precip processing system Supports a prototype 0 -6 h QPF with skill comparable to that of HPC forecasters 32

NMQ Precipitation Products Cool season: Number of HRAP grid boxes with radar-reference correlations ≥ 0. 50 and ≥ 0. 70 were 26% and 35% higher for NMQ than for PPS 33

Cool Season Effective Radar Coverage for Precipitation: Green: Areas where NMQ Precipitation improves Radar-gauge correlation From poor-marginal to useful (correlation > 0. 5) Blue: Converse 34

Median peak stage error (m) for 5 storm events on subbasins of the Tar River, North Carolina, 2003 -2006. Basins range in size from 2300 km 2 (ROKN 7) to 116 km 2 (SIMN 7). NMQ is from MRMS, HPE from operational NEXRAD PPS. 35

Assimilation of reflectivity (RUC DEV 13) enables prediction of higher rainfall amounts, which are only rarely generated by RUC without radar ingest (NCEP OPER). This includes amounts up to 2 inches in 12 hours. Weygandt, S. , and S. Benjamin, 2007: Radar reflectivity-based initialization of precipitation systems using a diabatic digital filter within the Rapid Update Cycle. Preprints, 18 th Conf. Num. Wea. Pred. , Park City, UT, Amer. Meteor. Soc. , 1 B. 7. 36

NMQ QC improves RTMA The MRMS Quality Control (QC) algorithm significantly reduces non-precipitation echoes – a potential improvement to RTMA 37

WDSS-II: Warning Decision Support System – Integrated Information http: //wdssii. nssl. noaa. gov • A suite of data ingest and quality control processes • Severe weather algorithms and applications • Uses multi-sensor observations • Products delivered to NCEP (SPC and AWC) and NWS FOs (OUN, FTW, TUL), CAPS, ESRL, NCAR and Environment Canada for operational use • An archive of automated Hail Swath and Rotation Track information 38

MRMS Severe Weather Algorithms • 3 D Reflectivity Mosaic Automated QC Neural Network VIL, Echo. Top, Composite, Isotherms Hail (Max Expected Size, Probability, Hail Swath) Kmeans/Segmotion Nowcast, Storm Tracking Storm Classification • 3 D Velocity Vortex Detection & Diagnosis Algorithm (from Azimuthal Shear) Multi-Doppler Wind Analysis • Multi-Sensor Algorithms Lightning (NLDN & LMA) Near Storm Environment (from RUC numerical model analyses) Satellite 39

MRMS Severe Weather Products “Rotation Tracks” “Hail Swaths” 40

NMQ/WDSS-II Computational Infrastructure The entire NMQ/WDSS-II processing system is composed of 2 Quad-Core AMD Opteron Processors (2. 9 GHz with 32 GB RAM) Linux servers from a single manufacturer (HP); Servers and server drives can quickly be configured, maintained and swapped out; Running Red. Hat 64 -bit OS; 41

Radar Data Ingest WSR-88 D • Ingest base level 2 data from 158 tri-agency (DOC, DOD, DOT) 10 cm WSR-88 D radars • Ingest base level data from 33 5 cm radars from Environmental Canada/NCDC • Ingest base level data from two TDWRs using direct connections • Ingest commercial radars in addition to experimental radar systems such as CASA and PAR Canadian Radars • System designed to accommodate increasing access to radar systems and radar networks TDWR 42

NMQ/WDSS-II Mosaic and Product Creation Radar Ingest WSR-88 D 4 D grids & products Precipitation Products Aviation Next. Gen Weather Products Severe Weather Products Canadian TDWR Multi-sensor Ingest Products are disseminated in Net. CDF, binary, AWIPS, N-AWIPS, GIS, and HRAP formats using the LDM protocol. 43