1 ESE Science Review [Meeting on Weather Forecasting] Eric A. Smith; NASA/Goddard Space Flight Center, Greenbelt, MD [ ; ; July 20, 2001; NASA Headquarters, Washington, DC GPM A Limited Perspective on Weather Forecast Science Problems Confronting GPM

2 TRMM 1-day coverage SSM/I Era TRMM Era EOS Era GPM Era

3 Projected Satellite Data Streams for GPM Era from Passive Microwave Radiometers & Precipitation Radars [at left are either actual (bold) or orthodox (paren) nodal crossing times (DN or AN) or non-sun-synch labels ] CY AMSR-E CMR-1 F20 NPOESS C1 F18 SSM/I SSMIS CMIS SSM/I SSMIS CMIS MSU  AMSU-A AMSR AMSR-FO GPM-1 (e.g., N-GPM) GPM Core (65 ∞ inc) ADEOS II PR/TMI DPR/ATMI TRMM (35 ∞  inc) DMSP F13 DMSP F16 NPOESS C3 DMSP F15 DMSP F17F19 NPOESS LITE-CMIS AQUA GPM-3 (e.g., E-GPM) GPM-2 (e.g., I-GPM) GCOM-B1 MEGHA-TROPIQUES (22 ∞  inc) MADRAS GPM-5 (partner needed) GPM-4 (partner needed) CMR-2 CMR-3 CMR -4 CMR-5 FY-3 TBD 0530DN 0130DN NSS 1030DN TBD NOAA-MNOAA-K CMIS MSU  AMSU-A 0730DN NOAA-LNOAA-N Potential Gap KEY carries preferred PMW frequencies carries alternate PMW frequencies GPM-1,-2,-3,-4,-5 are dedicated GPM drone satellites (e.g., N-GPM, I-GPM, E-GPM) NOAA-J (1430 AN ) (1730 AN ) (2030 AN ) (2330 AN ) (1030 DN ) (0230 DN ) NPOESS C2 NOAA-N ’  Continuous Geosynchronous Satellite Coverage by GOES E/W, METEOSAT/MSG, & GMS  (0830 DN ) (0530 DN ) NPP-ATMS Replacement Era 0915DN 0830DN

4 GPM Mission Reference Concept GPM Advanced Study Formulation

5 GPM Science Agenda

6 GPM Constellation Orbit Optimization

7 GPM Mission is Being Formulated within Context of GWEC with Main Science Objectives Focusing On: _____________________________________________________________________________________________ Improving flood hazard & basin-scale hydrological predictions -- through more frequent sampling and full-earth coverage of precipitation measurements Improving climate prediction -- through better understanding of water cycling and accompanying accelerations -- decelerations of atmospheric and surface branches of water cycle Improving weather forecasting -- through better methods of rainfall data assimilation and more accurate & precise measurements of instantaneous rainrates

8 GPM Era Coverage Drones GPM Era Coverage with 3 Inclined GPM Core, DMSP-F18, DMSP-F19, GCOM-B1, Megha-Tropiques, & Three 600-km 34 ∞, 84 ∞, 90 ∞ 3-hour Ground Trace

9 Tropical Cyclone Prediction Conundrum

10 TRMM Impact on Mesoscale Simulation of Super Typhoon Paka PAKA (8.9 ∞ N, ∞ E) [13 Dec-1997 / 0911 UTC] SSM/I 85 GHz TB GEOS with TRMM RR & TPW + Bogus Vortex (adjoint-based 4-D VAR) [13 Dec-1997/ 0900 UTC] RR(mm/3hr); LP; 850 hPa Wind GEOS without TRMM [13 Dec-1997 / 0900 UTC] RR(mm/3hr); SLP; 850 hPa Wind GEOS with TRMM RR & TPW [13 Dec-1997 / 0900 UTC] RR(mm/3hr); SLP; 850 hPa Wind [Pu & Tao, 2001: GSFC] 33 hr forecasts using PSU/NCAR MM5 model at 5-km horizontal resolution testing different initial conditions for time 12 Dec-1997 / 0000 UTC

11 Retrieval of Rainrate Vertical Structure Conundrum

12 Data Assimilation Experiments Based on Retrieved SSM/I & TRMM Rainrates Have Not Been Particularly Sensitive to Intensity of Rainrates Nor Have Made Use of Vertical Profile of Rainrate or Latent Heating

13 Cloud-Precipitation Continuum Conundrum

14 Impact of Rainfall Assimilation on GEOS Analysis Assimilation with TMI+SSM/I rainfall & TPW Precipitation verified against GPCP ____________________ TPW verified against Wentz ____________________ OLR verified against CERES/TRMM ____________________ IR Cld Forcing verified against CERES/TRMM Assimilation of satellite-based rainfall data improves clouds & TOA radiation, plus reduces state-dependent systematic errors in GEOS analysis OLR Error Std Dev verified against CERES for 1-, 5-, & 30-day Averaging Periods Control Run

15 Physical Initialization Under Complex Convective Parameterization Scheme Conundrum

16 Reverse Schemes for Convective Parameterizations Carrying Updrafts & Downdrafts Do Not Yield Unique Solution for Adjusting (nudging) Water Vapor Field Once Model-Observation Surface Rainfall Departures Are Determined

17 Control variables at time t 0 Control variables at time t Surface rainrate Forecast Model (PHYSICS) Interpolation at observation location MOIST PHYSICS (convection+ grid scale condensation) Increments of control variables at time t 0 Increments of control variables at time t Surface rainrate departure Adjoint of Forecast Model (PHYSICS) Adjoint of spatial interpolation Adjoint of MOIST PHYSICS 4DVAR Rainfall Data Assimilation Forward Backward

18 GPM Validation Strategy Tropical Continental Confidence sanity checks GPM Satellite Data Streams Continuous Synthesis ∑ error variances ∑ precip trends Calibration Mid-Lat Continental Tropical Oceanic Extratropical Baroclinic High Latitude Snow Research Quality Data Algorithm Improvements Research ∑ cloud macrophysics ∑ cloud microphysics ∑ cloud-radiation modeling FC Data Supersite Products II. GPM Supersites  Basic Rainfall Validation hi-lo res gauge/disdrometer networks polarametric Radar system  Accurate Physical Validation scientists & technicians staff data acquisition & computer facility meteorological sensor system upfacing multifreq radiometer system D o /DSD variability/vertical structure convective/stratiform partitioning III. GPM Field Campaigns  GPM Supersites cloud/ precip/radiation/dynamics processes  GPM Alg Problem/Bias Regions targeted to specific problems I. Basic Rainfall Validation  ∑  Raingauges/Radars new/existing gauge networks new/existing radar networks

19 Focused Field Campaigns Meteorology-Microphysics Aircraft GPM Core Satellite Radar/Radiometer Prototype Instruments Piloted UAVs 150 km Retrieval Error Synthesis Algorithm Improvement Guidance Validation Analysis Triple Gage Site (3 economy scientific gages) Single Disdrometer/ Triple Gage Site (1 high quality-Large Aperture/ 2 economy scientific gages) 150 km 100-Gage Site Lo-Res Domain Centered on Multi-parm-Radar 5 km 50-Gage Site Hi-Res Domain Center-Displaced with ∑ Uplooking Radiom/Radar System [10.7,19,22,37,85,150 GHz/14,35,95 GHz] ∑ 915 or 2835 MHz Doppler Radar Profiler ∑ Portable X-band Radar Data Acquisition- Analysis Facility DELIVERY Legend Multiparameter Radar Uplk Radiom/Radar 940 MHz Profiler Port X-band Radar Meteorological Tower Supersite Template Site Scientist (3) Technician (3)

20 ECMWF Requirements for GPM from Rainfall Assimilation Experience Spatial Resolution: Well-defined rain product spatial resolution (ECMWF-model will be going to 15 km forecast / 30 km assimilation resolutions) Sampling: Prefer “less often but more accurate” Error Considerations: Quantification of error in rain detection Quantification of retrieval errors/time-space biases Removal of inter-satellite retrieval errors Assessment of errors due to spatial/temporal sampling mismatch Plans at ECMWF: Evaluation of rainrate vs simplified radiance assimilation Improved estimation of humidity profile forecast errors

21 Final Comments Currently, NASA is only organization with wherewithal to bring online global observing system of precipitation & closely related data assimilation variables which could significantly improve weather forecasting through improvements in water-related components of numerical prediction models & associated data assimilation schemes. However: Better observations by themselves do not solve or resolve all standing problems in predictive modeling and thus ESE plan must resolve which groups and through what mechanisms model & data assimilation technique development will proceed to take advantage of current and future space measurements NASA intends to provide.