Progress of DCC Method in FY3A/MERSI and FY2

Presentation on theme: "Progress of DCC Method in FY3A/MERSI and FY2"— Presentation transcript:

1 Progress of DCC Method in FY3A/MERSI and FY2
Chen Lin, Xiuqing Hu, Ling Sun CMA GPRC, National Satellite Meteorology Center

2 OUTLINE Background DCC for Instrument Performance Monitoring of FY-3/MERSI DCC for Instrument Performance Monitoring of FY2D &FY2E Conclusions

3 Background WHY DCC Advantage Disadvantage No need navigation;
Many targets; No sharp BRDF; Suitable for both GEO and LEO Easy to achieve …… Disadvantage Effect of stratosphere aerosols ; Rely on Thermal IR Too many data ……

4 Radiative Transfer Model Simulation
SBDART model Surface: Sea water Background Tropospheric Aerosols : Oceanic 0.1; Background Stratospheric Aerosols: 0.02 Cloud Parameters: Ice cloud, Size 106um, Height 2-12km Ratio= The change of Reflectance with COD when COD>100, the ratio tend to 0

5 Aerosols Effect Tropospheric Aerosols: Oceanic with AOD 0.1,0,2,0.5
Stratospheric Aerosols: Background 0.1 and 0.2 Fresh Volcano Ash 0.1 and 0.2

6 effect the DCC Reflectance Angle Distribution Models
Angles Effect Different Angles will effect the DCC Reflectance Angle Distribution Models Should be Considered R， Anisotropic Factor Anisotropic Factor at SZA 15°（a）、25°（b）、35°（c）and 45°（d） over Ocean for Clear Sky of ice cloud with optical depth 50 ADMs based on CERES TRMM observations

7 DCC as IPM for FY3/MERSI Methodology
DCC identification: Thermal IR TBB <205K; Latitude: 15°S to15°N； Longitude：0-360°； Surface： Ocean; Uniformity test: TBB in 3*3 grid box all less than 205K; STD of TBB in 3*3 grid box <1k STD of Ref in 3*3 grid box <1.5% Angles: Sun Zenith Angle<30°；View Zenith Angle<40° Statistics Bin: 10d or 30d Using the Pre-Launch Calibration table to calculate the nominal reflectance

8 Spectral specification of MERSI bands
Center wl (mm) Width (mm) IFOV (m) NEDρ(%)/ NEDT (300K) Signal Dynamic Range 1 0.470 0.05 250 0.45 100% 2 0.550 0.4 3 0.650 4 0.865 5 11.25 2.5 0.54 K 330k 6 1.640 1000 0.08 90% 7 2.130 0.07 8 0.412 0.02 0.1 80% 9 0.443 10 0.490 11 0.520 12 0.565 13 14 0.685 15 0.765 16 17 0.905 0.10 18 0.940 19 0.980 20 1.030 4 channels IFOV 250m 2 shortwave IR channels (1640;2130) 4 channels central wavelength below 500nm(470nm 250m;412nm; 443nm;490nm) 3 water vapor channels

9 Comparison 10d and 30d bin 10d disperse

10 normalization 2σ 2σ :Double STD of DCC means to the fit curve

11 Sensitivity of TBB Threshold
SZA and VZA Threshold is 40°， Ref bin=0.01 Mode：max(pdf) Mean: average Medium:mid-value SZA和VZA的阈值为40度，bin为0.01，考虑了ADMs校正，不同的红外TBB阈值判识的情况下，得到PDF的峰值随判识阈值变化。 Mode的值要大于平均值及中值，因为PDF的形态存在偏态分布，因此平均值和中值要小于mode值。平均值的稳定性最差，随TBB阈值变化较明显。中值最稳定。

12 ADMs Effect Without ADMs With ADMs VZA 0-10;11-20;21-30,31-40
根据ADM分布图看出，对于角度在40°以内，是否经过ADMs有3~7%的差别。如果不经过校正，对于绝对辐射定标而言，就会引入3~7%的误差。虽然对于辐射定标跟踪来说影响应该不会很大，因为相当于是否都除以一个校正因子，但是对于其定标跟踪的稳定性也会起到一定的修正作用。 从经过ADMs校正前后，不同观测角度情况下的反射率PDF分布上来看，经过校正的反射PDF一致性要明显好于未校正，其PDF的形态和最大概率点都比较一致。具体数据见表格： 除了VZA在31-40°之间的mode值与别的角度下有差异，其余角度的mode值经过校正后就一致了。且无论Mode、Mean还是Medium不同角度的相对方差均表明经过ADMs校正后的值比未经过ADMs校正的方差均要小。说明ADMs的校正对于实现DCC的稳定性有一定的帮助。 VZA With ADMs Without ADMs Count Mode Mean Medium 0-10 0.760 0.704 0.666 0.831 0.772 0.706 30144 11-20 0.709 0.665 0.840 0.785 0.715 256436 21-30 0.779 564305 31-40 0.750 0.695 0.655 0.830 0.762 0.700 540100 0-40 0.701 0.774 STD 0.0060 0.0074 0.0069 0.0064 0.0111 0.0098

13 Non-Linear decay of MERSI
First Year Second Year Third Year Forth Year Total linear nonlinear STD linear STD nonlinear Rate% 5.24 3.47 2.44 2.45 13.87 14.01 0.091 0.078

14 The difference of 3 kinds of Method for monthly average
Total Nonlinear Mean Mode Medium STD nonlinear Rate% 14.3 14.82 14.4 0.078 0.082 Mean 和 medium得到的衰减率更接近。月平均结果与相应拟合线的标准差来看，Mean方式得到的标准差较小。但每个通道的情况可能会有所不同。

15 Degradation has a little
Blue Channels 11.25 14.3 470(250m) 20.81 443 Degradation （Mean %) Wavelength (nm) 490 37.22 412 Shorter the channels wavelength is, Degradation is faster Degradation has a little bit nonlinear effect with time Significant Degradation

16 Red Channels(1) 2.29 565 5.57 550(250m) Degradation (Mean %)
Wavelength (nm) 2.57 520

17 Red Channels(2) Seem to be a lit bit rise -1.73 685 -2.06 650
Degradation (%) Wavelength (nm) -2.69 650(250m) Seem to be a lit bit rise

18 Near Infrared Channels
1.48 865 -0.17 865(250m) Degradation (%) Wavelength (nm) 1030 -0.33 765 15.88 The most stable channel

19 Water Vapor Channels 7.78 980 8.99 940 Degradation (Mean %) Wavelength
5.17 905

20 WV wing

21 Shortwave Infrared Channels
Jump Jump

22 Total DegradationRate2 (%)
DCC Multi-Site Band 2σ/Mean (%) Total DegradationRate2 (%) Jul to Jul. 2009 Jul to Jul. 2010 Jul to Jul. 2011 Jul to Jun. 2012 Total Degra Rate2 (%) Multi-DCC 1 14.03 5.24 3.47 2.44 2.76 3.12 17.81 7.85 6.17 5.38 3.78 2 5.58 1.97 1.12 1.21 1.16 2.49 8.93 3.94 2.46 2.99 3.35 3 -2.26 -1.19 -0.5 -0.73 -0.56 2.31 -2.41 -1.45 -1.18 -0.37 -0.15 4 0.34 -1.91 0.73 0.04 0.67 2.08 -0.30 -3.00 -0.80 0.26 -0.64 8 35.88 9.98 8.54 8.03 9.61 5.05 37.30 15.40 10.11 8.02 1.42 9 19.67 6.34 4.55 4.04 5.93 3.68 23.45 10.19 7.26 6.55 10 11.03 3.82 2.88 2.04 2.2 2.70 15.35 6.72 4.71 4.32 11 2.62 -0.42 0.4 -0.91 0.15 2.68 11.79 6.01 3.49 3.88 9.17 12 1.07 0.64 0.11 0.12 2.03 5.88 2.92 2.21 3.42 13 -1.73 -1.3 -0.33 -0.67 -0.53 -1.29 -1.62 -1.89 -1.07 0.44 14 -1.44 -0.94 -0.41 -0.84 -0.65 2.12 -1.56 -1.60 -0.75 -0.12 15 1.59 0.96 0.32 0.06 0.02 2.67 1.77 1.04 0.22 0.37 0.18 16 -2.56 -0.1 -1.58 1.54 1.19 -0.83 1.84 17 4.57 -0.31 0.58 0.29 4.22 6.98 2.41 18 7.32 -0.52 1.58 1.48 14.95 14.22 6.9 19 6.96 1.32 1.34 5.09 9.97 3.01 20 13.15 -0.28 4.05 3.32 3.96 15.73 2.73 3.33 3.58 2.58 The difference between DCC and Multi-site in most Bands are smaller than 4% There are big differences in Band 10 and 11 band Center wl (mm) Width (mm) IFOV (m) NEDρ(%)/ NEDT (300K) Signal Dynamic Range 8 0.412 0.02 1000 0.1 80% 9 0.443 10 0.490 0.05 11 0.520 12 0.565 13 0.650 14 0.685 15 0.765

23 DCC as IPM for FY2 Methodology
DCC identification: Thermal IR TBB <205K; For the time of later August 2011, <203K Latitude: 20°S to20°N； Longitude：85°E to 125°E(FY2E)； 55°E to 95°E(FY2D)； Surface： Ocean; Uniformity test: TBB in 3*3 grid box all less than 205K/ <203K; STD of TBB in 3*3 grid box <1.5k STD of Ref in 3*3 grid box <1.5% Angles: 15°<Sun Zenith Angle<40°；View Zenith Angle<30° Statistics Bin: 30d For FY2E images time is between 5:00 and 8:00； For FY2D images time is between 3:00 and 6:00；

24 About 5% Degradation of FY2E in the past 2 years
Compare to FY2E, FY2D is quite stable Big difference of DCC reflectance between FY2D and FY2E is questionable

25 Conclusion CMA began a testing of DCC method for FY-3A/MERSI.
DCC method provide good consistent results for MERSI degradation with other method. It is a good reference for MERSI degradation monitoring. DCC can obtain more stable results for NIR water vapor bands with low seasonal oscillation. DCC method is ongoing for FY-2 visible band. DCC inter-calibrtaion for FY-2 visible band based on MODIS will be conducted in the near future (Following Dave Doelling ATBD). Uncertainty analysis of DCC will be also conducted.

26 Comments or Suggestions
Thanks! Comments or Suggestions