Sensor Web Observing System Simulator SWS Project üHigh

Sensor Web Observing System Simulator (SWS) Project üHigh Fidelity Weather Modeling üHigh Performance Computing üNetwork of Observing Systems September 9, 2008 Glenn J. Higgins (Northrop Grumman) Michael Seablom (NASA Goddard) 1

Agenda • Overview and Background • Architecture • Example Use Cases • Limited Simulation Example • Plans 2

Overview • Sensor Web Observing System Simulator (SWS) is an ongoing NASA project contractually supported by Northrop Grumman (NG) • Outgrowth of NASA Earth Science Technology Office (ESTO)funded activity leading to two major reports • Goal is to quantify impact of candidate changes to weather forecast system on weather forecast accuracy – Support trade studies prior to investment, leading to cost savings • SWS creates test-bed for simulating future weather forecasting concepts, e. g. , new sensors, targeted observing, etc. – – 3 Extension to climate monitoring observing systems desired NRC decadal missions highest priority Support Mission PI’s in defining mission requirements and impacts Computationally intensive and large data sets

Genesis of SWS • ESTO Weather Architecture Studies (2002† & 2004‡) • Existing technology can provide: • New technology will be needed to enable, especially if resources limited: • SWS project launched to provide simulation environment to test these ideas and ultimately facilitate trade studies – Fundamental improvement in predictive skill may be obtained by incorporation of an additional feedback between the forecast model and the observing system – More frequent data collection in or near regions where the forecast model predicts development of a significant feature (e. g. , GOES rapid scan mode) – – – Changing the observing mode of a satellite (power, slewing) Dynamic asset deployment (e. g. , UAVs) Communication between the model and the sensor Data collection in a more intelligent manner Cloud-free lines of sight assessment Higher density data collection in sensitive regions; lower density elsewhere Clausen, M. , Kalb, M. , Mc. Conaughy, G. , Muller, R. , Neeck, S. , Seablom, M. , Steiner, M. , 2002: Advanced Weather Prediction Technologies: NASA’s Contribution to the Operational Agencies, ESTO Technical Report. Higgins, G. , Kalb, M. , Lutz, R. , Mahoney, R. , Mauk, R. , Seablom, M. , Talabac, S. , 2004: Advanced Weather Prediction Technologies: Two-Way Interactive Sensor Web & Modeling System, ESTO Technical Report. 4

SWS High Level Architecture Needs to accommodate current system and future concepts 1 2 3 4 6 5

Some Potential Use Cases: Quantifying Impact of Proposed Changes to Forecasting System on Forecast Accuracy Doppler Lidar- 3 D Winds UAV measurements to improve hurricane track/ intensity forecasts Impact of different forecasts models Impact of new CONOPS 6 XOVWM- sea surface winds (XOVWM: Extended Ocean Wind Vector Winds Mission)

Example of Integrating SWS and Sensor Models § SWS + EVEREST enable end-to. NG EVEREST end sensor impact studies, including measurement accuracy-to-forecast accuracy Core SWS Components Measurement Uncertainties Of Other Sensors Scenario Definition Candidate Sensor Data Measure -ment Simulation Truth Data (Nature Run) 7 Data Assimilation Model GSI Forecast Model GEOS 5 Forecast Parameters Each forecast cycle Performance Analysis üSimulated Weather Data üReal Weather Models üSimulated Weather Data for New Sensors üHPC Forecast Quality Performance Metrics EVEREST: Environmental Product VErification and Remote Sensing Testbed

Evolution of Weather Forecast Predictive Skill- a metric for comparing TIME SERIES of monthly mean anomaly correlations for 5 -day forecasts of 500 h. Pa heights for various operational models (CDAS frozen as of 1995) - Northern Hemisphere “Anomaly Correlation” An expression of how well predicted anomalies correspond to observed anomalies One metric of predictive skill of weather forecasts Improvements in predictive skill over the past several decades have been gradual; the sensor web provides an opportunity for a “revolutionary” impact Source: Fanglin Yang, Environmental Modeling Center, National Centers for Environmental Prediction, NOAA 8

Use Case: Decadal Survey Mission 3 D Wind Lidar Global Wind Observing Sounder (GWOS) Life: 6 billion shots Source: Kakar, R. , Neeck, S. , Shaw, H. , Gentry, B. , Singh, U. , Kavaya, M. , Bajpayee, J. , 2007: An Overview of an Advanced Earth Science Mission Concept Study for a Global Wind Observing Sounder.

Application of Sensor Web Concepts • Simulation: Extend Mission Life via Power Modulation • Conserve power / extend instrument life by using aft shots only when there is “significant” disagreement between model first guess line-of-sight winds and winds measured by fore shots – Lidar engineers have recently suggested reduced duty cycles may increase laser lifetimes – Duty cycles that are on the order of 10 minutes “on” and 80 minutes “off” may be very beneficial to mission lifetime • Some combination of space/ground processing would be required – Requires engineering trades be performed for on-board processing, storage, power, weight, communications 10

Simulation Results Lidar data deleted when there is “adequate” agreement with the numerical model’s first guess wind fields Designed to simulate suppression of the aft shot of the lidar 11 Result: Nearly 30% of the lidar’s duty cycle may be reduced -- IF there is no discernible

Simulation Results (cont. ) 1 Results Anomaly Correlation versus Forecast hour Northern Hemisphere Full lidar set and targeted lidar set are nearly identical - indicating a reduced duty cycle may be possible Forecast Hour Southern Hemisphere Results in the Southern Hemisphere are more ambiguous; some indication of degradation due to targeting is evident Forecast Hour Impact of duty cycle reduction on forecast skill, 20 day assimilation with 5 -day forecasts launched at 00 z each day. Results represent an aggregate over all forecasts 12

Computational Requirements for the simulation functionality demonstration • Computer System: NCCS (NASA Center for Computational Sciences) Palm System (SGI Altix) • Number of processors used: 64 CPUs • Global Model Resolution: 1 x 1 deg 72 levels • Wall Clock: 10 Simulation Days/ Day • Number of Days Processed: 20 days • Data Requirements: – Input: 10 GB/simulation day – Output: 10+ GB/simulation day • Note: Higher resolution simulations over longer times periods are planned requiring more computational resources 13

Status and Plans • Improve user interface, automate production, increase scope of simulation capability • Conduct new simulations – – GOES-R (Geostationary Operational Environmental Satellite-R) GWOS (Global Wind Observing Sounder) XOVWM (Extended Ocean Wind Vector Winds Mission) PATH (Precipitation and All-Weather Temperature and Humidity) Mission • Integrate NG Everest sensor modeling capabilities through collaboration with NGIT and NGST, brings – NPOESS (National Polar-orbiting Operational Environmental Satellite System) – Other sensors 14

Q&A 15

Backup 16

Data Assimilation Today noitcello. C ata. D ”edo. M yevru. S“ demrofrep si sisylana ata. D. 2 si tsacerof laciremu. N. 3 detucexe ta detcelloc era snoitavresb. O. 1 & ecaps ni snoitacol ”enituor“ emit

Data Assimilation with Intelligent Sensor Webs Autonomous and On-Demand Targeting to Collect “Best” Observations 3. Numerical forecast is executed 4. Forecast error is estimated 2. Data analysis is performed 1. Observations are collected at specified locations in space & time Adaptive Targeting Automated / Manual Sensor Web Feedback Loop

Adapting SWS for other Decadal Missions • Decadal Missions have multiple purposes: – – – • Climate research and monitoring Health issues related to the environment Ecosystem research and monitoring Water resources monitoring Weather prediction and research The current SWS simulator design is aimed at modeling proposed changes to the forecast system and their impact on weather forecasting skill – It is not limited to changes in the observing system • 19 The application of the SWS to other focus areas of the decadal missions, such as climate monitoring, requires additional use cases to further drive the design