Validation of EOMAP Water Quality Products

Validation of EOMAP Water Quality products

contact: [email protected] www.eomap.com

Version: 20210223 This report is updated when new validation data becomes available For further information, please visit http://eoApp.eomap.com and download the actual whitepaper from the help section Content

Summary Methodology for satellite derived water quality MIP-EOMAP References Technology

Global Examples - Europe - North and South America - Asia - Africa - Australia

Assessable uncertainties | Harmonization level and long-term consistency: Comparable to in-situ methodologies of the satellite approach | Uncertainties in magnitudes: approx. 10 – 50 % worse in relation to in-situ methologies (reflecting the in-situ-related inconsistencies on multi-agency level) | Exceptions: Chlorophyll in CDOM rich lakes partly overestimated by factor 2-3. Unknown validity: Chlorophyll and absorption in turbid rivers, iron- and calcareous dominated waters Methodological | Methodology: In situ <> satellite (MIP-EOMAP) | Sampling location: point <> area integrated measurement differences | Sampling depth: various depth <> z90 light penetration depth | Intrinsic measures: various <> Absorption and scattering spectra | Uncertainty impact factors: various <> Recording conditions, inversion methodology Sampling methodology | Turbidity: Backscattering [1/m] at 500nm, calibrated to [NTU] (1 NTU = 0.0118 1/m) | Suspended Matter: Backscattering calibrated to [mg/l] (default: log10(TSM) = 1.02log10(TUR) -0.04 ) MIP satellite | Chlorophyll: Pigment absorption [1/m] at 440nm, calibrated to [µg/l] 1 µg/l=0.035 1/m | CDOM: Absorption of dissolved organic materials in [1/m] at 440nm | Other: Total organic absorption, total anorganic absorption in [1/m] at 440nm Sampling methodology | Turbidity: Side-scattering, calibrated to units of [NTU] or [FTU] | Suspended Matter: Gravimetric [mg/l], or turbidity calibrated to [mg/l] In-situ | Chlorophyll: HPLC, photometric, fluorometric, unit [µg/l] | CDOM: unit absorption [1/m]

Coverage | Global

Sampling rate & spatial | Daily: 500 - 250 m | Weekly: 30 - 10 m resolution | Up to daily on request: 5 - 1 m | ASTER (1999), Landsat 5 (1984-2013), Landsat 7 (1999), Landsat 8 (2013), | MERIS, MODIS Aqua/Terra (2002/1999), RapidEye (2009-2020), Satellite sensors | Sentinel-2A/B (2015/2017), Sentinel-3A/B (2016/2018), SPOT & | Worldview 2/3 (2009/2014) | Turbidity, Chlorophyll-a, Secchi Depth, Signal Depth, Total Suspended Matter, Product examples | Subsurface Reflectance, Remote Sensing Reflectance, Data processing | Standardized, physics based, automized sensor independent processing. Accounting for the adjacency and terrain altitude impact, fully | coupled and bidirectional atmospheric & in-water retrieval of harmonized water quality properties, reflecting for the full range of components | scattering and absorption in natural waters. detection of water, land, cloud, cloud shadows. Data workflow | Fast delivery: Between few hours and 2 days after acquisition, depending on sensor and continent. EOMAP processors are installed in | satellite data ground segments and cloud computing environments. A dedicated IT infrastructure ensures automated product archiving, components | aggregation, quality control, web access and dissemination

Remark: Actual sampling rate depends on regional conditions such as cloud over statistics Product validity is restricted to optical deep water bodies at minimum extend of 3-5 times the spatial resolution

Bresciani M., Giardino C., Stroppiana D., Dessena M.A., Buscarinu P., Cabras L, Schenk K., Heege T., Bernert H., Bazdanis G. & Tzimas A. (2019): Monitoring water quality in two dammed reservoirs from multispectral satellite data, European Journal of Remote Sensing, 10 pages, DOI: 10.1080/22797254.2019.1686956 Karle, N., Wolf, T., Heege, T., Schenk, K., Klinger, P., Schulz, K. (2019): Satellite Remote Sensing of Chlorophyll and Secchi Depth for Monitoring Lake Water Quality – A Validation Study. Processings for the SPIE remote sensing publication conf 10.Sept. 19, Strasbourg Dörnhöfer, K., Klinger, P., Heege, T., Oppelt, N. (2017): Multi-sensor satellite and in situ monitoring of phytoplankton development in a eutrophic-mesotrophic lake. Science of the Total Environment 612C (2018) pp. 1200-1214 DOI information: 10.1016/j.scitotenv.2017.08.219 Dörnhöfer, K., Göritz, A., Gege, P. , Pflug, B., Oppelt, N. (2016): Water Constituents and Water Depth retrieval from Sentinel-2A – A first evaluation in an Oligotrophic lake. Remote Sensing, 8, 941, doi:10.3390/rs8110941 GLaSS project 2016: D4.2 Validation report for nearby Lakes. Available upon request Eder E., Dörnhöfer K., Gege P., Schenk K., Klinger P., Wenzel J., Oppelt N., Gruber N. (2016): Analysis of mineral-rich suspended matter in glacial lakes using simulations and satellite data, Proc. ‘Living Planet Symposium 2016’, Prague, Czech Republic, 9–13 May 2016 (ESA SP-740, August 2016), p 1-6 Broszeit, A., 2015. Assessing long-term inland water quality using satellite imagery: A Feasibility and validation study of different lake types. MSc Thesis, Julius-Maximilian-University Würzburg, 96p Heege, T., Kiselev, V., Wettle, M., Hung N.N. (2014): Operational multi-sensor monitoring of turbidity for the entire Mekong Delta . Int. J. Remote Sensing, Special Issues Remote Sensing of the Mekong, Vol. 35 (8), pp. 2910-2926 Koponen S., Kallio, K., Pyhälahti, T., Attila, J., Piepponen, H., Keto, V., Stelzer, K, Schenk, K., Krah, S., Heege, T. (2013): Update for Report on FRESHMON data quality and data comparability. FRESHMON project report. Available upon request Koponen S., Kallio, K., Hommersom, A., Pitarch, J., Schenk, K., Krah, S., Heege, T. (2012): Report on FRESHMON data quality and data comparability. FRESHMON project report. Available upon request Hausknecht, P. (2010): Operational MODIS Satellite based water turbidity monitoring for dredging operations in Woodside. OGP Remote Sensing workshop at ESA, Sept 2015, DRIMS 5526810. Available at https://www.eomap.com/exchange/pdf/Hausknecht.pdf (16.11.2015)

Heege, T.; Schenk K., Wilhelm M.-L. (2019): Water Quality Information for Africa from Global Satellite Based Measurements: The Concept Behind the UNESCO World Water Quality Portal. P. 81 – 92. In: Embedding Space in African Society. The United Nations Sustainable Development Goals 2030 Supported by Space Applications. Editors: Froehlich, Annette (Ed.) Heege, T., Kelleher, D. 2018: Reducing economic risks in hydropower developments through independent satellite-based turbidity and sediment measurements in the river systems of Georgia. Proc. Hydro 2018 conference, Gdansk 15.-17.10.2018, pp15. Heege, T., Bergin M., Hartmann K. and Schenk K. (2016): Satellite Services for Coastal Applications. Chapter 18, pages 357 – 368. Ocean Solutions, Chapter 18, Dawn J. Wright, ed.; 2016; Earth Solutions, second edition; http://dx.doi.org/10.17128/9781589484603_18 Kiselev, V., Bulgarelli, B. and Heege, T., 2015. Sensor independent adjacency correction algorithm for coastal and inland water systems. Remote Sensing of Environment, 157: 85-95. Giardino, C., Bresciani, M., Cazzaniga, I., Schenk, K., Rieger, P., Braga, F., Matta, E., Brando, V.E., 2014. Evaluation of Multi-Resolution Satellite Sensors for Assessing Water Quality and Bottom Depth of Lake Garda. Sensors 14, 24116-24131; Doi:10.3390/s141224116 Heege, T. , Häse, C. , Bogner, A. and Pinnel, N. 2003. Airborne Multi-spectral Sensing in Shallow and Deep Waters. Backscatter p. 17-19, 1/2003 Kiselev, V.B., Roberti, L. and Perona, G., 1995. Finite-element algorithm for radiative transfer in vertically inhomogeneous media: numerical scheme and applications. Appl. Opt. , 34, 8460-8471. https://www.esri.com/training/assets/courses/576605f98cace5ed09c827a3/Oceans2_sample.pdf

160 Oligotrophic lake Näsijärvi 140 Mesotrophic lakes Vesijärvi 120 Peipsi Eutrophic lakes Littel Harris 100 Monroe Altmühlsee 80 Hypereutrophic lakes Apopka Beauclair 60 Satellite chl-a Satellite Dora

40 R^2 = 0.77 SD = 20.60 20 p = <0.0001 N = 316

0 0 20 40 60 80 100 120 140 160 Processor MIP version: 2015 Q3, Reference: Broszeit 2015 In-situ data kindly provided by: Syke Finnish Environment Institute, In situ chl-a Lake County Water Authority US/Florida, Water Supply Zürich, BOWIS / IGKB by LUBW, Bavarian Environment Agency (LfU)., Tartu Observatory Estland

MIP satellite CHL-a [µg/l] in situ CHL-a [µg/l] 100

10 Chlorophyll-a[µg/l]

Lake Peipsi 2013Altmühlsee 2014 Lake Dora 2010 Lake Dora 2007 Lake Näsijärvi 2005 Lake VesijärviLake Monroe 2009 2014 Lake ApopkaLake Beauclair 2010 2014Lake ApopkaLake Beauclair 2008 2007 Lake Constance 2011 Lake Little HarrisLake 2011 LohjanjärviLake Little 2002 Harris 2007 Processor MIP version: 2015 Q3, Reference: Broszeit 2015 Großer Brombachsee 2014 In-situ data kindly provided by: Syke Finnish Environment Institute, Lake County Water Authority US/Florida, Water Supply Zürich, BOWIS / IGKB by LUBW, Bavarian Environment Agency (LfU)., Tartu Observatory Estland

Location River Elbe, Germany Lake/river size 1.900 km² Time Period 2010 Parameter Turbidity Sensor Landsat 5 and 7 ETM+ Spatial Resolution 30m Validation data provided by Portal Tideelbe www.portal-tideelbe.de

insitu.15.min satellit 36 34 32 30 28 26 24 22 20 18 16

14 Turbidity

[ETU/NTU] 12 10 8 6 4 2 0

18.04 25.04 11.05 04.06 05.06 20.06 21.06 28.06 29.06 07.07 10.10 19.10 Station Pagensand 2010 Satellite data: processing MIP © EOMAP, source data: USGS for Landsat 5 and 7 ETM+ Processor MIP version: 2014 In situ data source: www.portal-tideelbe.de In-situ data kindly provided by: Portal Tideelbe, www.portal-tideelbe.de Reference: BAW contract

insitu 70 Landsat 5

[NTU] 30 Turbidity

00:00 01:30 03:00 04:30 06:00 07:30 09:00 10:30 12:00 13:30 15:00 16:30 18:00 19:30 21:00 22:30 04.06.2010 Steinriff

Satellite data: processing MIP © EOMAP, source data: USGS for Landsat 5 Processor MIP version: 2014 In situ data source: www.portal-tideelbe.de In-situ data kindly provided by: Portal Tideelbe, www.portal-tideelbe.de Reference: BAW contract

Location Lake Constance, Germany Lake/river size Approx. 530 km² Time Period 2003 - 2018 Parameter Chlorophyll-a, Total Suspended Matter Sensor MERIS, MODIS, Landsat 7 ETM+, Landsat 8, ASTER, SPOT Spatial Resolution 500m, 300m, 250m, 30m, 20m, 15m Stations FU, field campaign stations Validation data provided by IGKB – Bodensee-Wasserinformationssystem BOWIS, EAWAG, ISF Mean signal depth ~3-5 m In situ depth 0-20 m Reference Karle N., Wolf T., Heege T., Schenk K., Klinger P., Schulz, K. (2019): Satellite remote sensing of chlorophyll and Secchi depth for monitoring lake water quality: a validation study. doi: 10.1117/12.2533233. FRESHMON project (2010-2013): D54.3 Report on FRESHMON data quality and data comparability (available upon request) D54.3_2 Update Report on FRESHMON data quality and data comparability (available upon request)

Karle N., Wolf T., Heege T., Schenk K., Klinger P., Schulz, K. (2019): Satellite remote sensing of chlorophyll and Secchi depth for monitoring lake water quality: a validation study. doi: 10.1117/12.2533233.

Validation Lake Constance 5 R2 = 0.87 Data range: 2003 - 2009 SD = 0.30 p = <0.0001 Lake Constance - FU - MERIS timeseries 4 N = 14

3 10

MERIS, Landsat 7 TUR [NTU] TUR 7 Landsat MERIS, 1 8

0 0 1 2 3 4 5 In situ TUR [NTU]

4 TSM in[mg/l] TSM

Total SuspendedTotal Matter 2

0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 MERIS Landsat 7 ETM+ Landsat 8 in situ, based on Secchi depth zs using conversion formula TSM= 5.3*(1/ln zs-0.28) (according Heege 1998) in situ data © Lake Constance water information system BOWIS of the International Commission for the Protection of Lake Constance (IGKB) provided in context of FRESHMON and GLaSS project; processing: MIP © EOMAP; Satellite data: ESA for MERIS and USGS for Landsat 7 ETM+/ 8

Lake Constance - FU - MERIS timeseries

20 g/l]

µ 10

Chl-a[ Chlorophyll-a

0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 in situ MERIS Landsat 7 ETM+ Landsat 8 in situ data © Lake Constance water information system BOWIS of the International Commission for the Protection of Lake Constance (IGKB) provided in context of FRESHMON and GLaSS project; processing: MIP © EOMAP; Satellite data: ESA for MERIS and USGS for Landsat 7 ETM+/ 8

Atmospheric conditions: Varying haze and clouds In-water conditions: Fast changing turbidity near river mouth Rhine => Time-consistency important

Lake Constance Field Campaign 25.05.2012 14

TSMfilt 25.May 2012 TSMwisp 25.May 2012 12 SPOT20 09:25 UTC ASTER 10:33 UTC MODIS250 10:30 UTC 10 MODIS250 12:15 UTC SPOT pm0.5 hours MODIS (12:15) pm 0.5 hours

TSM [mg/l] TSM 6

0 StationsNr: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 15 16 17 18 19 20 21 22 23 24 25

07:00 07:30 08:00 08:30 09:00 09:30 10:00 10:30 11:00 11:30 12:00 12:30 13:00 13:30 14:00 14:30 15:00 15:30

Location Several lakes in Baden-Württemberg, Germany Time Period 2004 - 2012 Parameter Chlorophyll-a Sensor Landsat 7 Spatial Resolution 30m References Karle N., Wolf T., Heege T., Schenk K., Klinger P., Schulz, K. (2019): Satellite remote sensing of chlorophyll and Secchi depth for monitoring lake water quality: a validation study. doi: 10.1117/12.2533233. FRESHMON project (2010-2013): D54.3 Report on FRESHMON data quality and data comparability (available upon request) D54.3_2 Update Report on FRESHMON data quality and data comparability (available upon request)

BR (Lake FU (Lake Mindelsee Federsee Bad Waldsee Illmensee Schluchsee Titisee Constance) Constance) No. Satellite 30 14 11 10 10 23 8 10 Years Satellite 2004-2012 2004-2012 2004-2012 2004-2012 2004-2010 2005-2012 2004-2012 2004-2012 No. in situ 32 53 8 202 220 18 7 10 Years in situ 2002,2009,2010 2006-2010 1987-2010 2003-2011 2003-2011 2006,2012 2010 2007 Source ISF ISF seenprogramm.de ISF ISF ISF ISF ISF

In situ vs. satellite derived Upper CHLa Limit

80 CHLa max satellite L7 CHLa in situ 70

30 CHL[µg/l]

Titisee Schluchsee Bsee-BR Bsee-FU Mindelsee Illmensee Bad Waldsee Federsee -- Selected Lakes in Baden-Württemberg

Source: in situ ISF/LUBW, Seenprogramm.de, satellite data: processing EOMAP, raw data USGS Lansat 7

Location Lake Starnberg, Germany Lake/river size Approx. 56.4 km² Time Period 2015 Parameter Remote Sensing Reflectance Sensor Sentinel-2A Spatial Resolution 10m Stations 7 Field campaign stations Validation data provided by/ Dörnhöfer, K. et al. (2016): Reference Water Constituents and Water Depth retrieval from Sentinel-2A – A first evaluation in an Oligotrophic lake

Comparison of three different atmospheric corrections: Sen2Cor, ACOLITE and MIP, whereof MIP performed best (r = 0.987, RMSE = 0.002 sr-1)

Location Ammersee, Germany Lake/river size 46,6 km² Time Period 2008 Parameter Chlorophyll-a, Total Suspended Matter Sensor MERIS Spatial Resolution 300m Stations Deepest point Validation data provided by Bavarian Environment Agency (LfU) Mean signal depth ~1-3 m In situ depth 0-20 m, single steps see graph Reference FRESHMON project (2010-2013): D54.3 Report on FRESHMON data quality and data comparability (available upon request) D54.3_2 Update Report on FRESHMON data quality and data comparability (available upon request)

Ammersee 8

TSM [mg/l] TSM

0 2008 2009 2010 2011

Landsat 7 ETM+ in situ (LfU) (based on Secchi depth zs using conversion formula TSM= 2,65*(1/ln zs-0.28)) MERIS with quality: <33% 33-66% >66%

Satellite data: processing EOMAP, source data: USGS for Landsat 7ETM+, ESA for MERIS Processor MIP version: 2013 In situ data by the Bavarian Environment Agency (LfU) provided in context of FRESHMON project In-situ data kindly provided by: Bavarian Environment Agency (LfU). Reference: EU FRESHMON Project

Ammersee 2008 30

g/l]

Chl-a[

Chlorophyll-a 10

0 Jan Feb Mrz Apr Mai Jun Jul Aug Sep Okt Nov Dez in situ 2m in situ 4m in situ 6m in situ 8m in situ 10m in situ 0-20m MERIS Satellite data: processing MIP © EOMAP, source data: ESA for MERIS Processor MIP version: 2013 In situ data by the Bavarian Environment Agency (LfU) provided in context of FRESHMON project In-situ data kindly provided by: Bavarian Environment Agency (LfU). Reference: EU FRESHMON Project

Location Walchensee, Germany Lake/river size 16.27 km² Time Period 2008 Parameter Chlorophyll-a, Total Suspended Matter Sensor MERIS, Landsat 7 ETM+ Spatial Resolution 300m, 30m Stations Deepest point Validation data provided by Bavarian Environment Agency (LfU) Reference FRESHMON project (2010-2013): D54.3 Report on FRESHMON data quality and data comparability (available upon request) D54.3_2 Update Report on FRESHMON data quality and data comparability (available upon request)

Walchensee 2008 5

1 Total SuspendedTotal [mg/l] Matter

0 Feb Mrz Apr Mai Jun Jul Aug Sep Okt

MERIS Landsat 7ETM+ in situ (based on Secchi depth zs using conversion formula TSM=5,3*(1/ln zs-0.28))

Satellite data: processing MIP © EOMAP, source data: USGS for Landsat 7ETM+, ESA for MERIS Processor MIP version: 2013 In situ data by the Bavarian Environment Agency (LfU) provided in context of FRESHMON project In-situ data kindly provided by: Bavarian Environment Agency (LfU). Reference: EU FRESHMON Project

Walchensee 2008 20 insitu 0-20m 18 MERIS 16 Landsat 7 ETM+ insitu4m 14 Insitu6m Insitu8m

12 insitu10m g/l]

µ insitu12m

8 CHL-a[

0 Feb Mrz Apr Mai Jun Jul Aug Sep Okt

Satellite data: processing MIP © EOMAP, source data: USGS for Landsat 7ETM+, ESA for MERIS In situ data by the Bavarian Environment Agency (LfU) provided in context of FRESHMON project Processor MIP version: 2013 In-situ data kindly provided by: Bavarian Environment Agency (LfU). Reference: EU FRESHMON Project

Location Altmühlsee & Brombachsee, Germany Time Period 2000-2015 Parameter Chlorophyll-a, Secchi Depth, Signal Depth Sensor Landsat Spatial Resolution 30m Validation data provided by Bavarian Environment Agency (LfU)

Validation Altmühlsee 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 0,0

-0,5

-1,0

-1,5

Secchi depth [m] depth Secchi Lineare regression satellite: -2,0 Y = 45.33 + -0.0000187 * x n = 108 p-value = 0.11 -2,5 Lineare regression in situ: Y = 131.99 + -0.000054 * x n = 252 p-value = <0.0001 -3,0 in situ Landsat 5 TM Landsat 7 ETM+ Landsat 8 OLI Processor MIP version: 2015 Q3 linear regression in situ data linear regression satellite data In-situ data kindly provided by: Bavarian Environment Agency (LfU) Reference: Broszeit 2015

Validation Altmühlsee 320 300 Lineare regression satellite: Y = 7380.5 + -0.003 * x 280 n = 81 p-value = 0.01 260 Lineare regression in situ: Y = 10828.5 + -0.004 * x 240 n = 279 p-value = 0.002 220 200 180 160

140 Chl-a [µg/l] Chl-a 120 100 80 60 40 20 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

in situ Landsat 5 TM Landsat 7 ETM+ Landsat 8 OLI Processor MIP version: 2015 Q3 linear regression in situ data linear regression satellite data In-situ data kindly provided by: Bavarian Environment Agency (LfU) Reference: Broszeit 2015

Validation Großer Brombachsee 2008 2009 2010 2011 2012 2013 2014 2015 0

-1

-2

-3

-4

-5 secchi depth [m]/ depth secchi

satellite z90 [m] z90 satellite -6 In situ In -7

-8

-9 in situ Landsat 5 TM Landsat 7 ETM+ Landsat 8 OLI Processor MIP version: 2015 Q3 In-situ data kindly provided by: Bavarian Environment Agency (LfU) Reference: Broszeit 2015

Validation Großer Brombachsee 60

30 Chl-a [µg/l] Chl-a

0 2008 2009 2010 2011 2012 2013 2014 2015 Processor MIP version: 2015 Q3 in situ Landsat 5 TM Landsat 7 ETM+ Landsat 8 OLI In-situ data kindly provided by: Bavarian Environment Agency (LfU) Reference: Broszeit 2015

Location Kummerower See, Germany Time Period 2015 Parameter Chlorophyll-a Sensor MODIS, Landsat 7, Landsat 8, Sentinel-2a Spatial Resolution 10-500m Reference Dornhöfer K. et al. 2017

Reference: Dörnhöfer K., Klinger P., Heege T., Oppelt N.: Multi-sensor satellite and in situ monitoring of phytoplankton development in a eutrophic-mesotrophic lake. Sci Total Environ. 2018 Jan 15;612:1200-1214. doi: 10.1016/j.scitotenv.2017.08.219. Epub 2017 Sep 8. PMID: 28892864.

Comparison of Chlorophyll-a retrieved from in situ measurements and Sentinel-2A acquisitions. Vertical error bars indicate the standard deviation of a 5 × 5 pixel environment. Horizontal error bars represent standard deviation of in situ measurements at the sampling sites, the grey shaded area indicates the 48% uncertainty of in situ measurements.

Lake average satellite Chlorophyll-a and in-situ Chlorophyll-a between 1 July and 3 October 2015. Vertical bars indicate standard deviation (a). Lake average eoHAB and cyanobacteria fraction of biomass (LU-MV, 2015a) (b).

Location Lake Mulargia, Italy Lake/river size Approx. 12.5 km² Time Period 2013-2018 Parameter Chlorophyll-a, Turbidity, SST Sensor Sentinel-2A, Landsat 7, Landsat 8, WorldView-2 Spatial Resolution 10-30m Stations 1 Field campaign station (39° 36' 27,1395''‚ 9° 15' 12,1778'') Validation data provided by CNR (Claudia Giardino, Mariano Bresciani), ENAS (Ente Acque della Sardegna) Reference SPACE-O project http://www.space-o.eu/

Variations of Chlorophyll-a concentrations from satellite data (Landsat 7, Landsat 8 and Sentinel 2) and in situ data for Lake Mulargia, in correspondence of the ENAS station.

Variations of Turbidity from satellite data (Landsat 7, Landsat 8 and Sentinel 2 platforms) and in situ data for Lake Mulargia, in correspondence of the ENAS station.

June 2017 July 2017 August 2017

September 2017 October 2017 November 2017

Location Aposelemis dam in Crete, Greece Lake/river size Approx. 2 km² Time Period 2013-2018 Parameter Chlorophyll-a, Turbidity, SST Sensor Sentinel-2A, Landsat 7, Landsat 8 Spatial Resolution 10-30m Stations Station 1 Reference SPACE-O project (not yet published) http://www.space-o.eu/

0.03 0.03

0.02

) 0.02

)

r r

s y = 1.2803x

(

2 R² = 0.8331

s s

r RMSE = 0.0036 sr-1

r R 0.01 R y = 1.2339x 0.01 R² = 0.9137 RMSE = 0.0021 sr-1

0 0 0 0.01 0.02 0.03 0 0.01 0.02 0.03 Rrs WISP (sr-1) Rrs WISP (sr-1)

0.03 0.03

0.02

) 0.02

)

r r

s y = 1.2803x

(

2 R² = 0.8331

s s

r RMSE = 0.0036 sr-1

r R 0.01 R y = 1.2339x 0.01 R² = 0.9137 RMSE = 0.0021 sr-1

0 0 Figures left: 0 0.01 0.02 0.03 0 0.01 0.02 0.03 Scatterplot across common bands of Rrs values from WISP and from satellite observations (on the bottom figure: S2 for 2nd Rrs WISP (sr-1) Rrs WISP (sr-1) and 7th July, on top : L8 for data measured on 03/07/18).

) 10

(

e 8

l e

t 6

y t

i 4

r u

T 2

0 0 2 4 6 8 10 12 SPM in situ (g/L)

)

m (

l l

e 3 t

a Parameter RMSE Slope Intercept R2 Av. In situ Av. Sat

D 2 D S -1 1 SPM (gL )/TUR 1.05 0.92 0.81 0.89 6.90 6.62 0 SDD (m) 0.41 0.94 0.20 0.88 2.27 2.21 0 1 2 3 4 5 6 SDD in situ (m) Chl-a (mgm-3) 5.11 1.02 4.87 0.77 17.66 12.54

30 )

3 25

m /

g 20

(

t i

l 15

t a

S Table on top:

a -

l Results of the statistical analysis for evaluation the accuracy of satellite derived products (Lake Mulargia and Aposelemis dam). h

C 5

0 Figure left: 0 5 10 15 20 25 30 Scatterplot of water quality parameters from in situ measurements and from satellite observations; red diamonds are for Lake Mulargia, blue for 3 Chl-a in situ (mg/m ) Aposelemis dam.

Location Lake Zurich, Switzerland Lake/river size Approx. 88,66 km² Time Period 2006 - 2011 Parameter Chlorophyll-a Sensor MERIS Spatial Resolution 300m Stations SZHTH Validation data provided by Water Supply Zurich Reference FRESHMON project (2010-2013): D54.3 Report on FRESHMON data quality and data comparability (available upon request) D54.3_2 Update Report on FRESHMON data quality and data comparability (available upon request)

Lake Zurich - SZHTH - Timeseries 2006-2011 30

g/l]

Chl-a[

Chlorophyll-a 10

0 2006 2007 2008 2009 2010 2011 in situ 2.5m in situ 10m MERIS Landsat 7 ETM+ Satellite data: processing MIP © EOMAP, source data: USGS for Landsat 7ETM+, ESA for MERIS In situ data © Wasserversorgung Zürich 2012 provided in context of FRESHMON project Processor MIP version: 2013 In-situ data kindly provided by: Water Supply Zürich Reference: EU FRESHMON Project

Location Räterichsbodensee & Gelmersee, Switzerland Lake/river size Approx. 0.67 km² & 0.645 km² Time Period 2013 Parameter Remote Sensing Reflectance, Subsurface Reflectance Sensor Landsat 8 Spatial Resolution 30m Reference Eder E., Roettgers R., Damm A., Schenk K., Odermatt D & Wuest A. (2014): Remote sensing of particle mass concentration in Alpine reservoirs. In: Ocean Optics XXII conference. Portland, Maine, USA, 26-31 October 2014.

0.12 0.45 insitu 2013, Jul 15th insitu 2013, Jul 15th insitu 2013, Aug 14th insitu 2013, Aug 14th insitu 2013, Sep 2nd 0.4 insitu 2013, Sep 2nd 0.1 LS8 2013, Jul 16th LS8 2013, Jul 16th LS8 2013, Aug 17th LS8 2013, Aug 17th 0.35 LS8 2013, Sep 2nd LS8 2013, Sep 2nd

0.08 0.3 ] -1 0.25 0.06

R(0-) 0.2 Rrs(0-) [srRrs(0-)

0.04 0.15

0.1 0.02 0.05

0 0 300 400 500 600 700 800 900 1000 300 400 500 600 700 800 900 1000 nm nm

0.07 insitu 2013,Jul 5th insitu 2013,Jul 5th insitu 2013,Jul 25th insitu 2013,Jul 25th insitu 2013, Aug 15th insitu 2013, Aug 15th 0.06 insitu 2013, Aug 29th 0.25 insitu 2013, Aug 29th LS8 2013, Jul 7th LS8 2013, Jul 7th LS8 2013, Jul 16th LS8 2013, Jul 16th 0.05 LS8 2013, Aug 17th LS8 2013, Aug 17th

LS8 2013, Sep 2nd 0.2 LS8 2013, Sep 2nd ]

-1 0.04

0.15 R(0-)

0.03 Rrs(0-) [srRrs(0-)

0.1 0.02 in-situ measured data vs. satellite 0.05 retrieved values (2013): 0.01 top: Räterichsbodensee, bottom: Gelmersee, 0 0 300 400 500 600 700 800 900 1000 300 400 500 600 700 800 900 1000 nm nm in situ provided by Elisabeth Eder presented at Ocean Optics 2014

LS8 2013, Sep 2nd 0.2 LS8 2013, Sep 2nd ]

-1 0.04

0.15 R(0-)

0.03 Rrs(0-) [srRrs(0-)

0.1 0.02

0.05 0.01

Gelmersee 0 0 in-situ measured data vs. satellite 300 400 500 600 700 800 900 1000 300 400 500 600 700 800 900 1000 retrieved values (2013): nm nm in situ provided by Elisabeth Eder, presented at Ocean Optics 2014

0.08 0.3 ] -1 0.25 0.06

R(0-) 0.2 Rrs(0-) [srRrs(0-)

0.04 0.15

0.1 0.02 0.05

Räterichsbodensee 0 0 300 400 500 600 700 800 900 1000 300 400 500 600 700 800 900 1000 in-situ measured data vs. satellite nm nm retrieved values (2013): in situ provided by Elisabeth Eder, presented at Ocean Optics 2014

Location Po river delta, Italy Time Period 2012-2014 Parameter Turbidity Sensor MERIS, Landsat 7 ETM+, Landsat 8 Spatial Resolution 500m, 30m Stations PO_SEA01 Mean signal depth ~2-5 m

Po Delta Validation Consistency test of Landsat 7, Landsat 8 und MODIS Station: SEA_PO01 100

10 TUR [FTU] TUR

01.04.2012 01.10.2012 01.04.2013 01.10.2013 01.04.2014

Satellite derived (MODIS) Turbidity (TUR) [FTU] Satellite derived (LSAT 7) Turbidity (TUR) [FTU] Satellite data processing: MIP © EOMAP Satellite derived (LSAT 8) Turbidity (TUR) [FTU] satellite data source: NASA for MODIS (Aqua and Terra) and USGS for Landsat 7 ETM+/8 Processor MIP version: 2015 Q3 Reference: Commercial contract ISPRA

Location Lake Vänern, Sweden Lake/river size Approx. 5600 km² Time Period 2015 Parameter Remote Sensing Reflectance Sensor Sentinel-2A, Landsat 8 Spatial Resolution 10m, 30m Stations Field campaign stations Validation data provided by GLaSS consortium partners (Water Insight, Brockmann Consult. Brockmann Geomatics, Tartu Observatory, SYKE, CNR, EOMAP, STICHTING VU-VUMC) Reference GLaSS Project (2013-2016): D4.2 Validation report for Nearby Lakes (available upon request)

Landsat 8

Reflectance spectra of WISP-3 and Landsat 8 (OLI) using MIP in-situ and satellite (both 30th of August 2015)

Good spectral shape and reasonable intensity between satellite and in situ

Sentinel 2

Reflectance spectra of WISP-3, RAMSES and Sentinel 2-A MSI for Lake Vänern in-situ (30th of August 2015) and satellite (29th of August 2015)

Large differences between in situ measurements

Location Bay Tvärminne, Finland Time Period June 2012 Parameter Signal Depth Z90, Total Suspended Matter, Turbidity Sensor RapidEye Spatial Resolution 5m Validation data provided by Finnish Game and Fisheries Research Institute Reference FRESHMON project (2010-2013): D54.3 Report on FRESHMON data quality and data comparability (available upon request) D54.3_2 Update Report on FRESHMON data quality and data comparability (available upon request)

9 N = 42

8 R2 = 0.22

7 RMSE = 0.96

6 Y = 0.49*X +0.96

4 June 18

In situ turbidity (NTU) turbidity situ In 3 June 19 June 20 2 June 21 (1:1) 1 Trendline

0 0 1 2 3 4 5 6 7 8 9 10 Processor MIP version: 2013 In-situ data kindly provided by: Syke Finnish Environment Institute Rapid Eye turbidity (NTU) Reference: EU FRESHMON Project

Processor MIP version: 2013 In-situ data kindly provided by: Syke Finnish Environment Institute Reference: EU FRESHMON Project

8 N = 56

7 R2 = 0.71

6 RMSE = 0.9

5 Y = 1.1*X -0.04

4 June 18 3 June 19

In situ Secchi depth (m) Secchi depth situ In June 20 2 June 21 (1:1) 1 Trendline

0 0 1 2 3 4 5 6 7 8 Processor MIP version: 2013 Rapid Eye Z90 (m) * 2 In-situ data kindly provided by: Syke Finnish Environment Institute Reference: EU FRESHMON Project

Location Little Harris, Florida/US Time Period 1999-2015 Parameter Chlorophyll-a, Secchi Depth, Turbidity, Sensor Landsat Spatial Resolution 30m Validation data provided by http://www.wateratlas.usf.edu/

Validation Lake Little Harris In situ secchi depth (Station ID: 44677) vs. satellite derived secchi depth 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 0,0

-0,5 Secchi depth [m] depth Secchi -1,0

Processor MIP version: 2015 Q3 In-situ data kindly provided by: Lake County -1,5 Water Authority, USF Water Institute (2015) in situ Landsat 7 ETM+ Landsat 8 OLI Reference: Broszeit 2015, http://www.seminole.wateratlas.usf.edu/water resourcesearch.aspx, 30.9.2015

Validation Lake Little Harris In situ turbidity (Station ID: 44677) vs. satellite derived turbidity 30

15 TUR[NTU] 10

0 Processor MIP version: 2015 Q3 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 In-situ data kindly provided by: Lake County Water Authority, USF Water Institute (2015) in situ Landsat 7 Landsat 8 OLI Reference: Broszeit 2015, http://www.seminole.wateratlas.usf.edu/water resourcesearch.aspx, 30.9.2015

Validation Lake Little Harris) In situ chl-a (Station ID: 44677) vs. satellite derived chl-a 100

Chl-a [µg/l] Chl-a 40

0 Processor MIP version: 2015 Q3 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 In-situ data kindly provided by: Lake County Water Authority, USF Water Institute (2015) in situ Landsat 7 ETM+ Landsat 8 OLI Reference: Broszeit 2015, http://www.seminole.wateratlas.usf.edu/water resourcesearch.aspx, 30.9.2015

Turbidity Chlorophyll-a

Validation Lake Little Harris Validation Lake Little Harris 25 80 2 Data range: 2003 - 2011 R2 = 0.73 Data range: 2003 - 2011 R = 0.63 SD = 1.19 70 SD = 8.32 p = <0.0001 p = <0.0001 20 N = 49 N = 43 60

50 15

10 30 Landsat 7 & 8 chl-a [µg/l] chl-a 8 & 7 Landsat Landsat 7 & 8 TUR [NTU] TUR 8 & 7 Landsat 20 5 10

0 0 0 5 10 15 20 25 0 10 20 30 40 50 60 70 80 In situ TUR [NTU] In situ chl-a [µg/l]

Processor MIP version: 2015 Q3 In-situ data kindly provided by: Lake County Water Authority, USF Water Institute (2015) Reference: Broszeit 2015, http://www.seminole.wateratlas.usf.edu/waterresourcesearch.aspx, 30.9.2015

Location Beauclair, Florida/US Time Period 2005-2014 Parameter Chlorophyll-a Sensor Landsat Spatial Resolution 30m Validation data provided by http://www.wateratlas.usf.edu/

Validation Lake Beauclair In situ chl-a (Station ID: 45143) vs. satellite derived chl-a 240

220

200

180

160

140

120

Chl-a [µg/l] Chl-a 100

0 Processor MIP version: 2015 Q3 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 In-situ data kindly provided by: Lake County Water Authority, USF Water Institute (2015) in situ Landsat 7 ETM+ Reference: Broszeit 2015, http://www.seminole.wateratlas.usf.edu/water resourcesearch.aspx, 30.9.2015

Location Lake Monroe, Florida/US Time Period 1999-2014 Parameter Chlorophyll-a, Turbidity Sensor Landsat Spatial Resolution 30m Validation data provided by http://www.wateratlas.usf.edu/

Turbidity Chlorophyll-a

Validation Lake Monroe Validation Lake Monroe 50 80 2 Data range: 2010 - 2014 2 R = 0.52 R = 0.51 Data range: 2010 - 2014 SD = 0.15 70 SD = 8.93 p = <0.0001 p = <0.0001 40 N = 25 N = 33 60

50 30

20 30 Landsat 7 & 8 chl-a [µg/l] chl-a 8 & 7 Landsat

Landsat 7 & 8 TUR [NTU] TUR 8 & 7 Landsat 20 10 10

0 0 0 10 20 30 40 50 0 10 20 30 40 50 60 70 80 In situ TUR [NTU] In situ chl-a [µg/l]

Location Apopka Lake, Florida/US Time Period 2000-2015 Parameter Chlorophyll-a, Turbidity Sensor Landsat Spatial Resolution 30m Validation data provided by http://www.wateratlas.usf.edu/

Turbidity Chlorophyll-a

Validation Lake Apopka Validation Lake Apopka 100 200 2 R = 0.45 Data range: 2005 - 2014 R2 = 0.44 Data range: 2005 - 2014 90 SD = 10.10 180 SD = 20.06 p = <0.0001 p = >0.0001 80 N = 74 160 N = 66

70 140

60 120

50 100

40 80

30 [µg/l] chl-a 7 Landsat 60 Landsat 7 TUR [NTU] TUR 7 Landsat

20 40

10 20

0 0 0 10 20 30 40 50 60 70 80 90 100 0 20 40 60 80 100 120 140 160 180 200 In situ TUR [NTU] In situ chl-a [µg/l]

Location Tohopekaliga Lake, Florida/US Time Period 2008-2014 Parameter Chlorophyll-a Sensor Landsat Spatial Resolution 30m Validation data provided by https://lakewatch.ifas.ufl.edu/

Tohopekaliga Lake Station: Tohopekaliga-Middle 100

[µg/l] 40

Chlorophyll-a 30

0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

in situ Landsat 7 ETM+ (lake polygon) Landsat 8 (lake polygon) Processor MIP version: 2015 Q3 Satellite data: processing MIP © EOMAP, source data: USGS for Landsat 7ETM+ and Landsat 8 In-situ data kindly provided by: Lake County Water Authority, USF Water In situ data © Florida Lakewatch (http://lakewatch.ifas.ufl.edu/) Institute (2015) Reference: Broszeit 2015, http://www.seminole.wateratlas.usf.edu/waterresourcesearch.aspx, 30.9.2015

Location Lake Titicaca, Bolivia/Peru Lake/river size ~ 1000 km² Time Period 2007-2010 Parameter Chlorophyll-a, Total Suspended Matter / Turbidity Sensor MODIS Spatial Resolution 250m Station Central station 123 Entre Islas Soto y Taquile Reference Eoworld project ESA – Worldbank Partnership Report: 80-85

Central Lake Titicaca Seasonal trend of Chlorophyll a

12 Coordinate: UTM E 448272 N 8269015, +/- 1200m Central station 123 Entre Islas Soto y Taquile MERIS data 2003 - Sept. 2011, EOMAP processor MIP 10 satellite retrieved concentrations In situ campaign 26.5.-30.5.2003

6 Chlorophyll[µg/l]a 4

0 2 3 4 5 6 7 8 9 10 11 12 Month

Processor MIP version: 2011 Reference: Commercial contract Woodside Energy. http://www.eomap.com/exchange/pdf/Hausknecht.pdf

Location Mekong, Vietnam Time Period June 2012 Parameter Turbidity Sensor MODIS, Landsat, RapidEye Spatial Resolution 250 m, 30 m, 5 m Validation data provided by SiWRR and SRHMC Vietnam Hydromod and GFZ Germany Reference Heege et al. 2014

Reference: Heege, T., Kiselev, V., Wettle, M., Hung N.N. (2014): Operational multi-sensor monitoring of turbidity for the entire Mekong Delta. Int. J. Remote Sensing, Special Issues Remote Sensing of the Mekong, Vol. 35 (8), pp. 2910-2926

Can Tho / Basac river comparison

100

In-situ data (SRHMC) Turbidity [NTU] Turbidity In-situ data (GFZ CT3) 10 In-situ data (Hydromod) Satellite derived data MODIS monthly median Processing: © EOMAP, satellite data: USGS for Landsat 7, NASA for MODIS, Blackbridge for RapidEye LANDSAT 7 (single records) in situ data provided by SRHMC, GFZ, Hydromod RapidEye (single records) MIP version 2012 2007 2008 2009 2010 2011 2012 Reference: Heege, T., Kiselev, V., Wettle, M., Hung N.N. (2014): Operational multi-sensor monitoring of turbidity for the entire Mekong Delta . Int. J. Remote Sensing, Special Issues Remote Sensing of the Mekong, Vol. 35 (8), pp. 2910-2926

MIP version 2012 30 40 Can Tho and Basac River Can Tho and Basac River 24 Febr. 2012 35 25.Febr. 2012 25 In situ In situ Rapideye 11:12 ICT LANDSAT ETM 10:08 ICT 30 SPOT 10:16 ICT ASTER 10:31 ICT 20 25

15 20

15 Turbidity[NTU] 10 Turbidity[NTU]

5 5

0 0 09:00 09:30 10:00 10:30 11:00 11:30 12:00 12:30 09:00 09:30 10:00 10:30 11:00 11:30 12:00 12:30 Indochina time ITC Indochina time ICT

Time coincidence of Reference: Heege, T., Kiselev, V., Wettle, M., Hung N.N. (2014): Operational multi-sensor monitoring of turbidity for the in-situ and satellite entire Mekong Delta . Int. J. Remote Sensing, Special Issues Remote Sensing of the Mekong, Vol. 35 (8), pp. 2910-2926

Location Assuan dam, Egypt Lake/river size Approx. 5248 km² Time Period 2016 Parameter Chlorophyll-a, Turbidity Sensor Sentinel-2A, Landsat 7, Landsat 8 Spatial Resolution 10-30m Stations 1 stations Reference UNESCO project (not published yet)

Assuan Station 1 10

5 Sentinel-2A Landsat 8

4 Landsat 7 Turbidity NTU in Turbidity

0 12.12.15 31.1.16 21.3.16 10.5.16 29.6.16 18.8.16 7.10.16 26.11.16 15.1.17 Sentinel-2A sunglint in summer Date months

Assuan Station 1 20

a a µg/l in - 10 Sentinel-2A Landsat 8

8 Landsat 7 Chlorophyll 6

0 12.12.15 31.1.16 21.3.16 10.5.16 29.6.16 18.8.16 7.10.16 26.11.16 15.1.17 Sentinel-2A sunglint in summer Date months

Location Lake Volta, Ghana Lake/river size Approx. km² Time Period 2016 Parameter Chlorophyll-a, Turbidity Sensor Sentinel-2A Spatial Resolution 10m Stations Station GHA00006 Reference Hydrology TEP

Lake Volta Station GHA00006

Do not use outside H-TEP project 20

10 Sentinel-2A

Turbidity NTU in Turbidity 8 in situ

0 26.10.2016 25.112016 Only two suitable scences, Date others haze, sunglint or clouds over stations

Lake Volta Station GHA00006

Do not use outside H-TEP project 40

a a µg/l in -

20 Sentinel-2A in situ

Chlorophyll 15

0 26.10.2016 25.11.2016 Only two suitable scences, Date others haze, sunglint or clouds over stations

Location Northwest-Australia Lake/river size ~ 1000 km² Time Period 2007-2010 Parameter Suspended Matter, Turbidity Sensor MODIS Spatial Resolution 250m Reference Woodside Energy presentation: http://www.eomap.com/exchange/pdf/Hausknecht.pdf

Operational satellite based water quality monitoring service: EWS and MIP system 400 + data sets since Oct. 2007, Validation for: MODIS 250m turbidity service

Hausknecht, P. (2010): Operational MODIS Satellite based water turbidity monitoring for dredging operations in Woodside. OGP Remote Sensing workshop at ESA, DRIMS 5526810.

Processor MIP version: 2007 In-situ data kindly provided by: Woodside Energy Reference: Commercial contract Woodside Energy