Wesley J. Moses., Washington, D.C., USA.

Transcription

1 Wesley J. Moses, Washington, D.C., USA.

2 Sensor Characteristics 2 Spatial Resolution Spectral Resolution Signal-to-Noise Ratio Temporal Resolution

3 Spatial Resolution 3 What is the dominant spatial scale of variability of optically discernible biophysical features in coastal and inland waters?

4 Baltic Sea Landsat th Aug 2015

5 5

6 6

7 7

8 8

9 9

10 30 m 10

11 Sub-pixel Coefficient of Variation 30 m True Color 90 m 240 m 11 < m 750 m 990 m > 0.3

12 Spatial Variation Study 12 Along-track data of IOPs and AOPs from in-water, shipboard, airborne, and spaceborne platforms Data Track Observations Segment (Defined by GSD) n: no. of segments k: no. of observations in a segment Sub-pixel Variation, 2 * [SI ] median GSD min 1000 m Bias correction (Sokal, R. R. and J. F. Rohlf Biometry: The Principles and Practice of Statistics in Biological Research)

13 Data 13

14 14

15 Transition GSD (GSD t ) 15 Transition Region Logarithmic Fit GSD t (= e 5.52 = 250 m) Region of moderate change in CV a Region of steep increase in CV a Methods: Log-Log (LL) Method Slope Percentile (SP) Method For Details: Moses, W. J., Ackleson, S. G., Hair, J. W., Hostetler, C. A. and Miller, W. D. (2016). Spatial scales of optical variability in the coastal ocean: Implications for remote sensing and in situ sampling, Journal of Geophysical Research: Oceans, 121(6): , doi: /2016JC

16 CV vs. GSD 16 CV a /[CV a ] 1000 CV a /[CV a ] 1000

17 CV vs. GSD 17 CV a /[CV a ] 1000 CV a /[CV a ] 1000

18 Platform & Instrument In-water; ac-s Shipboard; ac-s Airborne; lidar Airborne; AVIRIS Airborne; CASI Spaceborne; HICO Transect Long Island Sound Parameter LL Method GSD t (m) SP Method [a p (450)] LIS [c p (650)] LIS Savannah New York [a p (450)] SN, [c p (650)] SN None None San Diego Coast [a p (450)] SD, [c p (650)] SD None None Virginia Coast [K sys (532)] VC [b b (532)] VC Mid-Atlantic Bight [b b (532)] MAB [K sys (532)] MAB Georges Bank and [b b (532)] GBM Gulf of Maine [K sys (532)] GBM 200* 200* [R rs_b ] SFB, [R rs_b/g ] SFB San Francisco Bay [R rs_g ] SFB [R rs_r ] SFB Monterey Bay 1 Monterey Bay 2 Monterey Bay Chesapeake Bay [R rs_b ] A-MB1, [R rs_g ] A-MB1, [R rs_r ] A-MB [R rs_b/g ] A-MB [R rs_b ] A-MB [R rs_g ] A-MB2 None None [R rs_r ] A-MB2, [R rs_b/g ] A-MB [R rs_b ] C-MB [R rs_g ] C-MB None 150 [R rs_r ] C-MB [R rs_b/g ] C-MB None None [R rs_b ] CB, [R rs_g ] CB, [R rs_r ] CB None None [R rs_b/g ] CB Spaceborne; Landsat-8 Baltic Sea [R rs_b ] BS, [R rs_g ] BS, [R rs_r ] BS, [R rs_b/g ] BS None None

19 Results 19 (for coastal waters) Average GSD t (LL): 173 m Average GSD t (SP): 217 m ~ 200 m GSD t generally smaller for regions closer to the shore (dominated by scattering processes) Does not imply that 200 m is sufficient; it simply means that beyond 200 m there is a significant loss in the ability to capture spatial variability

20 20 Cédric Jamet and Hubert Loisel Spatial resolution of ~200 m for capturing most changes in coastal Rrs

21 Spectral Resolution 21 What spectral resolution is required to resolve spectral features in complex inland and coastal waters?

22 Ryan Vandermeulen, Antonio Mannino, and Aimee Neeley 22

23 Signal-to-Noise Ratio (SNR) 23 Trade-off amongst Spatial Resolution Spectral Resolution SNR

24 Trade-off 24

25 Trade-off 25 SNR 150

26 What does SNR 150 mean for retrievals? 26 A single case study Add noise to Rrs spectrum at SNR = 700 and 850 Estimate chl-a for both cases and compare to the estimate from noiseless Rrs spectrum to determine the uncertainty due to noise, U SNR Effects of SNR on atmospheric correction not considered here U SNR=700 = 0.17 ± 4.5% U SNR=850 = 0.15 ± 3.7%

27 Impact of spatial aggregation on R rs (blue) Pahlevan et al. (2017) 27 Spatial averaging can only qualitatively improve the product quality. Even after spatial averaging, the OLI-derived products are of higher quality (due to its better SNRs). Pahlevan, N., Sarkar, S., Franz, B. A., He, J. Sentinel-2 MultiSpectral Instrument (MSI) data processing for aquatic science applications: Demonstrations and preliminary validations. Submitted to Remote Sensing of Environment

28 Signal-to-Noise Ratio (SNR) 28 Atmospheric correction is a major contributor of uncertainties in retrievals Lin Qi, Zhongping Lee, Chuanmin Hu, and Menghua Wang once SNR(NIR) is above 600:1, an SNR(vis) better than 400:1 will not make a significant reduction in product uncertainties

29 TSS (gm -3 ) Temporal Resolution 29 TSS concentration in Sacramento-San Joaquin River from SPOT Take 5 Data Christiana Ade et al. (see poster)

30 Synergistic Use of Satellite Data 30 Lake Kinneret (Israel) HICO Image 11 Mar Mar 2013 MODIS R645 Phycoerythrin 0.02 Phycocyanin Rrs (sr -1 ) Wavelength (nm) The presence of phycoerythrin and phycocyanin was confirmed by lab analysis of water samples

31 Acknowledgments 31 Help with data and analysis for the Spatial Resolution Study: Dr. Steven Ackleson (Naval Research Laboratory) Dr. Dave Miller (Naval Research Laboratory) Dr. Emmanuel Boss (University of Maine) Dr. Chris Hostetler (NASA Langley Research Center) Dr. Jonathan Hair (NASA Langley Research Center)

32 32 Contact: