Vertical profiles of aerosols in the lowest 300m - What we can see in CALIPSO observations and COSMO-MUSCAT model -

Transcription

1 Chart 1 Vertical profiles of aerosols in the lowest 300m - What we can see in CALIPSO observations and COSMO-MUSCAT model - Diana Mancera Supervisors DLR: Dr. Marion Schroedter-Homscheidt Dr. Lars Klüser Supervisors University of Oldenburg: Dr. Detlev Heinemann Professor Jürgen Parisi. DNICAST 2 nd End-user Workshop German Aerospace Center Oberpfaffenhofen, Germany December 1, 2015

2 Chart 2 OUTLINE I. The problem II. Approaches for aerosol vertical profiling III. Datasets IV. Merging processes V. Preliminary results VI. Sources of error in satellite data VII. Oncoming steps VIII. Conclusions

3 Chart 3 > Diana Mancera> October Short I. The Problem Distance emitter-receiver? m m m km Large CSP technologies. LFC Linear Fresnel Collector. PTC Parabolic Through Collector. DC Dish collector. - Long path (light) Extinction! Tower (central receiver) plant Ivanpah Solar Power Plant (392MW ). California, U.S. Landsat 8 satellite image. December 25, 2013.

4 Chart 4 > Diana Mancera> October I. The Problem Double counting of extinction? Apparently more sensitive to vertical distribution of aerosols because geometrical arrangement but. dispatchable!

5 Chart 5 I. THE PROBLEM The questions How much irradiance does the receiver actually get? Energy yield calculations: what is currently used? Raytracing software based on models assumptions: Standard atmospheric / limited scenarios Exponentially decrease of aerosol density with z: true for every place? Sometimes: specific inputs but parameters not available from campaigns. How could we improve the evaluation? Describing the height resolved aerosol profiles What is the aerosol contribution in the relevant altitude of a solar tower plant (~300m)? How much aerosol variation is there from place to place? (global mapping)

6 Chart 6 >> Diana Mancera> October II. Approaches to acquire vertical profiles Measurements & Estimations GROUND: AIRBORNE: NUMERICAL SATELLITE: LIDAR nadir-viewing MODELS: (Light Detection LIDAR /samplingbased +time And Ranging-LASER+Telescope +no incomplete flexibility: instruments backwards-long overlap lens) term +allows +no data(site +nearly incomplete vertical global assessment) coverage profiling overlap real time and?long -geographically forwards(forecasting -no real term time data for restricted operational capability) purposes -geographically -expensive -computing?uncertainties demanding restricted -incomplete -require?long term high data overlap accuracy in the at reproducing proximity of the processes device Aerosols? LIDAR Incomplete overlap function. Source: Navas-Guzmán et. at 2011

7 Chart 7 - Data available from Three instruments onboard: 1. Wide field camera(wfc) 2. Imaging infrared radiometer (IIR) 3. CALIOP(Cloud-aerosol Lidar with Orthogonal Polarization): - LASER pulses at 532nm and 1064nm: interaction with several kind of particles. - Generation of vertical curtains (profiles of clouds and aerosols). - Vertical resolution (L.II products): 60m (from -0,5km to 8,2km altitude). - Horizontal resolution (L. II 5km along the orbit (no grid!) (from -0,5km to 8,2km altitude). III. DATASETS Satellite - CALIPSO 60m 60m Source image: GES DISC NASA Source video: NASA (YouTube) 5km

8 Chart 8 III. DATASETS Satellite - CALIPSO - Repeat Orbital Cycle: 16 days (sun-synchronous orbit) - Space between 2 succesive equator ( 2h difference): 24,7 2741,7km - 15 orbits in 24h. Combined grid of CALIPSO overpasses in 16 days. Zoom North America Combined global grid of CALIPSO overpasses in 16 days Combined grid day/night in 24h 24,7 Source: CALIPSO (NASA) Website

9 Heigh(z) Chart 9 III. DATASETS Model COSMO-MUSCAT Longitude(x) Z 40 (x,y) Z 1 (x,y) Sigma-height coordinate system Partial AOD (z40) Partial AOD (z1) - TROPOS model - Modeling System. Coupling of: COSMO (nonhydrostatic limited-area atmospheric prediction model) MUSCAT (Chemistry Transport Model) Tegen 2002 (Dust emission scheme) - Vertical coordinate system: hybrid sigma-height 40 layers (up to 18km) - Horizontal: 0,25 (lat and lon) - Time intervals: 3 to 6 hours. - Aerosols included: purely dust 0,25 0,25 0,25 0,25 MUSCAT datasets (altitudes + extinction coefficient)

10 Chart 10 IV. MERGING PROCESS CALIOP data (public) Individual orbit files Level II Product Domain: Global (limited to orbits) Format: HDF Horizontal resolution: 5km resolution (following orbit) Vertical resolution: 60m CALIPSO DATA COSMO MUSCAT DATASET COSMO MUSCAT AOD files + Z files Format: GRIB 3 6 hour interval files Horizontal resolution : 0,25 Vertical resolution: Sigma height levels Domain:Regional(North Africa Europe) IDL routines -Satellite orbit screening -Cloud screenning -Calculation of AOD - Two comparable datasets: information of vertical aerosol distributions. - Possibly reflecting seasonal patterns.

11 Chart 11 IV. MERGING PROCESS Merge Model - Satellite - Different numbers of vertical layers: MUSCAT: Variable altitudes for each gridbox and heterogenous bins length. CALIPSO: fixed altitudes - Usage of cloud filters to obtain reliability from CALIPSO data (reduction of valid data).

12 Chart 12 V. PRELIMINARY RESULTS CALIPSO AOD 300m (all aerosols) January 2007 Geographical distribution AOD 300m Statistical distribution AOD 300m Uncertainties AOD 300m (%)

13 Chart 13 V. PRELIMINARY RESULTS AOD total - January 2007 Geographical distribution AOD total Statistical distribution AOD total Uncertainties AOD total (%)

14 Chart 14 V. PRELIMINARY RESULTS Fraction (AOD 300m /AOD total) - January 2007 Geographical distribution AOD total Statistical distributionaod total Uncertainties AOD total (%)

15 Chart 15 CALIPSO 532nm (all aerosols) V. PRELIMINARY RESULTS Comparison - AOD 300m January 2007 COSMO MUSCAT (alone all month) CALIPSO 532nm (only dust) COSMO MUSCAT (dust)

16 Chart 16 CALIPSO 532nm (all aerosols) V. PRELIMINARY RESULTS Comparison - AOD total January 2007 COSMO MUSCAT (alone all month) CALIPSO 532nm (only dust) COSMO MUSCAT (dust)

17 Chart 17 V. PRELIMINARY RESULTS Comparison - Fraction (AOD300m /AOD total) - January 2007 CALIPSO 532nm (all aerosols) COSMO MUSCAT (alone all month) CALIPSO 532nm (only dust) COSMO MUSCAT (dust)

18 Chart 18 > Vortrag > Autor Dokumentname > Datum V. PRELIMINARY RESULTS COSMO-MUSCAT - AOD 300m January Daytime January February March April May June

19 Chart 19 > Vortrag > Autor Dokumentname > Datum V. PRELIMINARY RESULTS COSMO-MUSCAT - AOD 300m January Daytime July August September October November December

20 Chart 20 > Vortrag > Autor Dokumentname > Datum V. PRELIMINARY RESULTS COSMO-MUSCAT Fraction (AOD300m /AOD total) January Daytime January February March April May June

21 Chart 21 > Vortrag > Autor Dokumentname > Datum V. PRELIMINARY RESULTS COSMO-MUSCAT Fraction: AOD300m /AOD total January Daytime July August September October November December

22 Chart 22 > Vortrag > Autor Dokumentname > Datum

23 CALCULATION OF UNCERTAINTY AT AVERAGING AODs (MONTHLY) PER GRIDBOX V. PRELIMINARY RESULTS Uncertainties CALIPSO DEVIATION FOR A MEAN IN MY APPROACH (Different quantities with the different true values. Not unique distribution) Based on NASA image of CALIFORNIA. Landsat 8 (OLI). October 2015

24 Chart 24 VI. SOURCES OF ERROR CALIPSO - Progressive attenuation of backscattered light - Changes in Signal Noise ratio - Sensititvity limitations: - Clear Air might not be clear air. - Sensitivity thresholds depending on aerosol type, altitude and day or night conditions and between the two channels. - Difficult to find a standard value to account for the aerosol background. Winker et. al Detection thresholds in terms of ext. coef for CALIOP 532nm. Upper: smoke. Bottom: marine.

25 Chart 25 VI. SOURCES OF ERROR CALIPSO - Missing information of opaque layers (signal totally attenuated) - Sometimes difficulties to retrieve true layer base Underestimation. - From time to time presence of large positive and negative excursions close to the surface Reported layer base True layer base Drawbacks at recognizing the true layer base CALIPSO in lowest aerosol layers

26 Chart 26 VII. ONCOMING STEPS Analysis SENSITIVITIES: - Cloudy vs. Cloudfree scenarios. - Mixed terrains (including steep regions) vs. Flater areas. - Screening supported by CALIPSO quality scores vs. No quality flags. CASE STUDY: - Libradtran (radiative transfer calculation software). - Comparison with ground measurements. PSA (transmittance).

27 Chart 27 VII. CONCLUSIONS State of the Art - Need of improving the capabilities of estimation and forecast of CSP plants energy yield - No information at all about vertical aerosol distribution in the lowest layer Expected achievements - Climatology datasets with understood limitations (different sources of data) - We will know wheater the dataset is meaningful for that purpose or not. -If it is the method and required routines are ready to be applied to further timeseries datasets. Possible future developments - - Validation of the two datasets with the help of airborne measurements. - Implementation of refinements to correct CALIPSO defficiencies in the lowest layers. -Feedback of particular needs to providers of measurements.

28 Chart 28 Thank you! Your questions

29 Chart 29 > Vortrag > Autor Dokumentname > Datum REFERENCES - Winker et.al The global 3-D distribution of tropospheric aerosols as characterized by CALIOP. - Winker et. al Overview of the CALIPSO mission and CALIOP data processing algorithms. - Goddard Earth Sciences Data and Information Services Center. GES DISC NASA. - Video CALIPSO: 3TM