INFORM project overview and status

Transcription

1 INFORM project overview and status Ils Reusen, VITO Els Knaeps, VITO and the INFORM consortium

2 Improved monitoring and forecasting of ecological status of European INland waters by combining Future earth ObseRvation data and Models

3 Why focus on inland waters? Fishing, recreation, water supply, transport, waste disposal, irrigation, increased pressures on EU inland waters asks for sustainable water management Monitoring of inland water quality required by EU Water Framework Directive (2000/60/EC) EU Habitats Directive (92/43/EEC) EU Shellfish Waters Directive (2006/113/EC) EU Drinking Water Directive (98/83/EC) EU Bathing Water Directive (2006/7/EC) EU Nitrates Directive (91/676/EEC) EU Urban Waste Water... Environmental Impact Assessments (e.g. by dredging companies)

4 Blueprint to Safeguard Europe's Water Resources (COM/2012/673) mentions Copernicus THE STATUS OF EU WATERS IS NOT DOING WELL ENOUGH! The Water Information System (WISE) will benefit from the development of INSPIRE, SEIS and Copernicus and from current water research works under FP7 and those to be conducted under H2020

5 Copernicus Copernicus = The European Earth Observation Programme Copernicus services address six main thematic areas:» Land Monitoring (operational)» Marine Monitoring (pre-operational phase)» Atmosphere Monitoring (pre-operational phase)» Emergency Management (operational)» Security (development phase)» Climate Change (development phase) All are provided free of charge to users

6 White Paper Copernicus Inland Water Services Discussion document to extend Copernicus Land Monitoring Service portfolio with inland water services demonstrated in FP7 Space projects Proposed Copernicus inland water services» Irrigation Water Abstraction Monitoring and Control Service» Pan-European Inland Water Quality Monitoring Service» Water scarcity and drought monitoring and forecasting Services» Inland water quantity monitoring service

7 Earth Observation for monitoring inland waters? Today: underutilized complexity and variability of these inland waters lack of adequate analysis methods lack of adequate low-cost EO data lack of uncertainty estimates Report GEO inland and coastal Water Quality Algorithm Workshop, Washington DC, May 2009: There is a lack of appropriate/dedicated satellite sensors for nearshore coastal and inland water quality applications.

8 Earth Observation & biogeochemical models? Assimilation of EO products into biogeochemical models allows for analysis of the cause-effect relationships governing a status change, forecast the response to pressures and evaluate different management actions. the future lies in the combined utilization of in situ data, remote sensing, and modeling. Tiffany A.H. Moisan, Shubha Sathyendranath and Heather A. Bouman (2012). Ocean Color Remote Sensing of Phytoplankton Functional Types, Remote Sensing of Biomass - Principles and Applications, Temilola Fatoyinbo (Ed.), ISBN: , InTech, Available from:

9 EU FP7-SPACE project INFORM Collaborative project - THEME [SPA ] [Remote sensing methods] Start date: 1/1/2014 Duration: 48 months 9 beneficiaries from 7 EU Member States Requested EU contribution: Grant agreement n

10 Main concept To develop novel user-driven products for inland water quality monitoring by using new innovative methods integrated into models which fully exploit the capabilities of upcoming Earth Observation missions (Sentinel-2, Sentinel-3, EnMAP and PRISMA)

11 Sentinel-2 Sentinel-2A launch readiness: 30 April 2015 Global revisit time: 5 days with 2 satellites MSI (Multi Spectral Instrument)» 13 spectral bands: 443 nm 2190 nm (including 3 bands for atmospheric corrections)» Spectral resolution: 15 nm 180 nm» Spatial resolution: 10 m, 20 m and 60 m» Swath: 290 km ESA P. Carrill

12

13 Sentinel-3 Sentinel-3A launch readiness: end of day global coverage OLCI (Ocean and Land Colour Instrument)» Swath width: 1270 km, with 5 tilted cameras» Spatial sampling: 300 m (full resolution mode)» Spectral range: 21 bands [ ] μm ESA PJ. Huart

14 Normalized OLCI SRFs, bands 1 to 21, plotted versus wavelength [nm] from C. Pelloquin, J. Nieke, SENTINEL-3 OLCI AND SLSTR SIMULATED SPECTRAL RESPONSE FUNCTIONS (S3-TN-ESA-PL-316) OLCI spectral bands = MERIS heritage+additional bands: Oa1 (400 nm): aerosol correction, improved water constituents retrieval Oa9 (673,75 nm): improved fluorescence retrieval and smile correction Oa14 (764,375 nm): atmospheric correction Oa15 (767,5 nm): cloud top pressure, fluorescence over land Oa20 (940 nm): water vapour absorption, atmospheric/aerosol correction Oa21 (1020 nm): atmospheric/aerosol correction

15 EnMAP» Expected launch date: 2017» Hyperspectral» Spectral range from 420 nm to 1000 nm (89 VNIR bands-8.1 nm FWHM) and from 900 nm to 2450 nm (155 SWIR bands-12.5 nm FWHM)» Swath width 30 km» Spatial resolution of 30 m x 30 m» Off-nadir (30 ) pointing feature for fast target revisit (4 days)

16 PRISMA» Expected launch date: 2017» Hyperspectral» Spatial resolution: m (Hyp) / m (PAN)» Swath width: km» Spectral range: µm (Hyp) / µm (PAN)» Continuous coverage of spectral ranges with 10 nm bands

17 APEX airborne hyperspectral imaging sensor for Simulation Calibration Validation of satellite sensors/products

18 INFORM consortium Participant organisation name VLAAMSE INSTELLING VOOR TECHNOLOGISCH ONDERZOEK N.V. CONSIGLIO NAZIONALE DELLE RICERCHE Participant short name Country VITO - Coordinator BELGIUM Ils Reusen, Els Knaeps, Sindy Sterckx, Liesbeth De Keukelaere, Dries Raymaekers, CNR ITALY Mariano Bresciani, Claudia Giardino, EOMAP GmbH & Co.KG EOMAP GERMANY Karin Schenk, Philip Klinger, Thomas Heege, THE UNIVERSITY OF STIRLING INSTITUT ROYAL DES SCIENCES NATURELLES DE BELGIQUE U STIRLING UK Peter Hunter, Andrew Tyler, Evangelos Spyrakos RBINS BELGIUM Dimitry Van der Zande, Kevin Ruddick, STICHTING DELTARES Deltares THE NETHERLANDS Miguel Dionisio Pires, PLYMOUTH MARINE LABORATORY MAGYAR TUDOMANYOS AKADEMIA OKOLOGIAI KUTATOKOZPONT KLAIPEDOS UNIVERSITETAS PML UK Giorgio Dall Olmo, Steve Groom + Stefan Simis, MTA OK HUNGARY Matyas Presing, KLAIPEDOS UNIVERSITETAS LITHUANIA Arturas Razinkovas- Baziukas,

19 INFORM Steering Advisory Board Members Tasks (SAB) Dr. Tiit Kutser, Remote Sensing and Marine Optics Department, Estonian Marine Institute, University of Tartu, Estonia Dr. Stewart Bernard, CSIR-NRE (Centre of High Performance Computing), South-Africa Dr. Vittorio Brando, CNR-IREA To provide recommendations at the SAB01 meeting (January 2014) To formulate scientific comments on the INFORM progress and to provide recommendations at SAB02 meeting (Mid-term, January 2016)

20 INFORM End-User Advisory Board Members (EUAB) Marc Sas/Boudewijn Decrop, International Marine and Dredging Consultants (IMDC), Belgium Marco Bartoli, Expert ecologist, University of Parma, Life Sciences Department Ute Menke, advisor Network Water, Rijkswaterstaat, the Netherlands István Kóbor head of laboratory, Central-Transdanubian Water Directorate, Hungary Geoff Phillips/Bill Brierley, Research, Monitoring and Innovation. Environmental Agency (EA) for England & Wales Alfred Johny Wüest, EAWAG, aquatic research institute, Switzerland Algirdas Stankevičius, Head of the Marine Research Department of the Ministry of Environment, Lithuania = COPERNICUS USER FORUM member Thomas Wolf, Environmental Agency of Baden-Wuerttemberg (LUBW), Germany Tasks To provide user requirements for INFORM developments at the EUAB01 (March 2014) and EUAB02 (Mid-term, January 2016) To attend the INFORM EUAB03 results uptake workshop (December 2017)

21 European approach Site Country Characteristic Lake Balaton Kis Balaton Hungary Largest shallow lake in Central Europe, meso-oligotrophic Water Protection System, hypereutrophic Curonian lagoon Lithuania Hypereutrophic coastal lagoon Lakes Mantua Italy Small and shallow artificial eutrophic basins Lagoon of Venice Italy Turbid coastal lagoon Lake Constance Germany, Meso-oligotrophic lake Switzerland, Austria Gironde river France Highly turbid river Scheldt river Belgium Highly turbid river Lake Windermere UK Mesotrophic lake Loch Lomond UK Warm, monomictic basin. Oligotrophic northern basin, mesotrophic southern basin Loch Leven UK polymictic, nonstratifying and eutrophic shallow lake Ijsselmeer The Netherlands Eutrophic lake, largest freshwater lake area in Northwestern Europe Markermeer is a turbid lake. +Lake Geneva, Switzerland

22 INFORM concept (detail) Properties of upcoming EO sensors (Sentinel-2, Sentinel-3, EnMAP, PRISMA) Improved spatial resolution Increased spectral coverage to shorter and longer wavelengths Improved spectral resolution Innovative analysis methods and improved atmospheric correction New/improved products Atttenuation and euphotic depth TSM and turbidity Yellow matter Phytoplanktion functional types Stratification Macrophytes Phytoplankton primary production Sun-induced chlorofyll fluorescence Improved modelling End-users End-users

23 WP objectives WP1 Management (VITO) Legal management Financial management Administrative management WP2 Scientific coordination (VITO) Scientific coordination of the project WP3 End-user interaction (CNR) To explore the end-user requirements in terms of water quality products To stimulate project results uptake by the end-users and industry WP4 Data gathering (VITO) To inventory existing data, identify data gaps To acquire new (in-situ, APEX hyperspectral and satellite) data» Development Campaign 2014» Testing Campaign 2016

24 WP objectives WP5 Algorithm development and validation (U STIRLING) Development and validation of EO products, and estimation of their uncertainty for WP6» Atmospheric correction (RBINS)» Attenuation and euphotic depth (RBINS)» TSM and turbidity (VITO)» Yellow matter (PML)» Phytoplankton functional types (CNR)» Stratification (EOMAP)» Macrophytes (CNR)» Phytoplankton primary production (U STIRLING)» Sun-induced chlorophyll fluorescence (U STIRLING)

25 WP objectives WP6 EO-model integration (Deltares) Integration of Earth Observation (EO) & In-Situ (IS) data and Water Quality (WQ) modelling WP7 Demonstration (EOMAP) To demonstrate to end-users» the INFORM prototype algorithms applied to new satellite sensors and» the added value of INFORM EO products for WQ model validation and forecasting To test the compliance of INFORM EO products with end-user requirements

26 WP objectives WP8 Dissemination (VITO) To disseminate the project objectives, progress and results To raise the awareness of the INFORM project To give recommendations for future satellite missions To organise a results uptake workshop

27 Interdependency of Work Packages

28 WP3: End-user interaction (Leader: CNR)

29 Kick-off end-user requirements EUAB end-user requirements formulated at the EUAB01 meeting, March 2014, Venice: General conclusion: the benefits that harmonized MULTI-TEMPORAL AND SPATIAL information derived from satellite images can give with respect to the traditional in-situ monitoring techniques based on point measurements was pointed out as the most important improvement compared to their current practices. In addition following requirements were formulated:» TEMPORAL AND SPATIAL RESOLUTION: Monthly temporal frequency of EO data, with a spatial resolution of 100 m. Exceptions are TSM, Turbidity and Chl-a maps which are required daily.

30 Kick-off end-user requirements» ACCURACY: Associated information about the quality of pixel values; robust algorithms with reference to literature or algorithm theoretical basis document (ATBD).» CONSISTENCY: Consistency between products derived from different sensors; a robust atmospheric correction with reference to literature or ATBD.» TAXONOMY: A standardized taxonomy (e.g. parameters names, measurement units, legend, color code) is received as a prerequisite for a harmonized EU-wide inland water quality monitoring.» ACCESSIBILITY: Easily accessible data and downloadable preferably by Web Map Service (WMS); training is requested.

31 WP4: Data gathering (Leader: VITO)

32 INFORM Development campaign - Balaton 2014 Lake Balaton and Kis Balaton wetland (Hungary) Data acquisition window: 7-28 July 2014 In-situ measurements (optical properties and water constituents) concurrent with satellite (Landsat8-OLI and HICO) and airborne hyperspectral (APEX) acquisitions

33 Balaton Balaton Limnological Institute Lake Balaton Kis Balaton Marcali Reservoir

34 Balaton - Characteristics Largest lake in central Europe Very shallow and well-mixed High mineral sediment loads (dolomitic mineralogy) Four distinct basins varying from mesotrophic to eutrophic Kis Balaton is hypertrophic Historically high nutrient loads but recent improvements in water quality Surface area 592 km 2 Catchment area 5772 km 2 Length 78 km Mean (max) width 9.1 (15) km Mean (max) depth 3.2 (11) m Water volume 1861 million m 3 Retention time Shoreline length 3-8 years 235 km

35 INFORM Development campaign - Balaton 2014 Quicklooks Landsat8-OLI acquisition Data available from the U.S. Geological Survey

36

37 INFORM Development campaign - Balaton 2014 Quicklooks HICO acquisition Collaboration with Evangelos Spyrakos, U STIRLING Data available from NRL The U.S. Naval Research Laboratory OSU Oregon State University

38 Lake Balaton Hungary Lake Balaton Hungary

39 INFORM Development campaign - Balaton 2014» APEX acquisitions and in-situ measurements»

40 INFORM Development campaign - Balaton 2014 Date EO Data acquisition In-situ sampling stations APEX Landsat-8 HICO USTIR CNR VITO ALL Comments 09/07/2014 X X /07/ /07/ Instrument inter-comparison + reference ground targets 16/07/2014 X 4 5 (+3 KB) Kis Balaton macrophyte measurements 17/07/2014 X APEX flights aborted 18/07/ /07/2014 X 4 4 (+2 KB) 10 X Kis Balaton macrophyte measurements + reference ground targets Water samples taken at 7 stations. Underway transects with radiometers 21/07/2014 X /07/ /07/2014 X X 6 6 TOTAL STATIONS SAMPLED

41 In-situ optics U STIRLING» Wetlabs AC-S: (size fractioned) spectral absorption and attenuation» Wetlabs Eco-BB3: spectral backscattering» CTD: temperature, salinity, depth» Trio Satlantic HYPEROCRs: subsurface irradiance reflectance» Trio Satlanctic HyperSAS and trio TriOS RAMSES: downwelling irradiance, skylight irradiance, total surface radiance for water-leaving reflectance» (in lab) TriOS OSCAR PSICAM: spectral absorption

42 In-situ optics CNR» Wetlabs AC-9 and Hobi Labs Hydroscat-6: spectral absorption and attenuation» Cyclops-6 fluorometers: phytoplankton pigments (Chla, PC, PE), CDOM fluorescence (+temperature and depth)» ASD FieldSpec FR and WISP-3: subsurface irradiance reflectance and remote sensing reflectance» ASD FieldSpec FR, Spectrascan, WISP-3: macrophytes reflectance VITO» ASD FieldSpec FR: remote sensing reflectance» WetLabs ECO-BB3: spectral backscattering

43 Water sample analysis Chla, PC, HPLC (pigments), particulate absorption (PABS) + flow cytometry TSM, CDOM, POC, DOC, phytoplankton cell counts Primary production Particle size distribution, mycosporine-like amino acids Macrophytes: dry weight biomass, pigment and nutrient analysis

44 Left: CIMEL CE318 - atmospheric measurements Middle: ASD FieldSpec FR - water reflectance measurements Right: Wetlabs AC-S, Wetlabs BB3, Wetlabs AC-9, Hobi Labs Hydroscat, Cyclops-6 fluorometers intercomparison of IOP measurements

45 Left: WetLabs AC-S and ECO-BB3 absorption and backscatter Middle: HYPERSAS and RAMSES downwelling irradiance, skylight radiance, total surface radiance Right: In-situ campaign leader Peter Hunter with Evangelos Spyrakos (U STIRLING)

46 Left: Filtering on the USTIR boat for pigments and particulate absorption. Middle: Filtering in the BLI lab for total suspended matter Right: Preparing samples in the BLI lab for dissolved organic carbon analysis

47 WP5: Algorithms development and validation (Leader: U STIRLING)

48 WP 5.1: Atmospheric correction Leader: RBINS Rationale Major source of uncertainity for EO products AC is very challenging for inland waters due e.g. to altitude, land adjacency and complex aerosols Objective To develop an atmospheric correction algorithm for Sentinel-2, Sentinel-3 and EnMAP/PRISMA for inland waters taking TOA radiance data and various auxiliary data as input and providing BOA water reflectance data as output

49 SIMEC adjacency correction Sterckx et al., RSE, in press

50 WP 5.2: Light attenuation & euphotic depth Leader: RBINS Rationale Key input for primary production and other ecological models Existing algorithms not suited to inland waters and need adaption to hyperspectral sensors Objective To develop/adapt algorithms for Sentinel-2, EnMAP/PRISMA and APEX taking water reflectance data as input and providing outputs for spectral and PAR diffuse attenuation coefficients (Kd, KdPAR) and euphotic depth (Ze)

51

52 WP 5.3: TSM & Turbidity Leader: VITO Rationale Key measure of water quality; cal/val of sediment transport and other ecosystem models Exploit new sensors, especially SWIR bands for high TSM Objective To develop/adapt algorithms for Sentinel-2, EnMAP/PRISMA and APEX taking water reflectance data as input and providing outputs for total suspended matter (TSM) concentration and turbidity (TUR).

53 Remote sensing reflectance 0.06 Varying Total Suspended matter concentration (mg/m3) [Credit: RBINS/VITO] Wavelength (nm)

54 [Credit: RBINS]

55 WP 5.4: Yellow matter Leader: PML Rationale YM = sum of absorption by CDOM and non-algal particles CDOM linked to DOC Major influence on short wavelength light availability Expoit new hyperspectral data products, including UV region Objective Develop and validate a UV-visible algorithm for yellow matter absorption that decomposes total absorption into pure water, pigments, and yellow substances

56 Dutch lakes data set Figure PML? + explanation Yellow matter absorption dominant but rarely isolated in UV-A region -> requires decomposition approach

57 WP 5.5: Phytoplankton functional types Leader: CNR Rationale Relative abundance of PFTs (or size classes) important to ecosystem function Some toxic bloom-forming species cyanobacteria pose risks to animal and human health also driver for WFD Objective To develop/adapt and validate algorithms for Sentinel-2, EnMAP/PRISMA and APEX taking water reflectance data and IOPs as input and providing outputs of Chla, secondary pigments, and size classes.

58 PE PC Chl-a [Credit: CNR]

59 PC in Esthwaite Water (UK) mapped using airborne AISA hyperspectral data [Hunter et al. 2010]

60 WP 5.6: Stratification Leader: EOMAP Rationale Lakes often have pronounced vertical gradients in dissolved and particulate material due to stratification Currently, methods provide no information on depth distribution Objective Feasibility study to derive information about vertical gradients of TSM using various satellite sensors.

61 MODIS 250m channels nm MODIS MODIS 500m channels nm MODIS 500 m channels look deeper than MODIS 250 m due to the incorporation of shorter wavelengths Shallow view 250m Applicable also to other sensors with different band combinations [Credit: EOMAP] Deep view 500m

62 WP 5.7: Macrophytes Leader: CNR Rationale Macrophytes fulfill important functional roles in lake ecosystems Biological quality element under EU WFD High spatial variability and coexistence of different species require high spatial resolution imagery Objective To develop/adapt classification approach for mapping different groups of macrophyte (emerged and submerged) and evaluate the biomass and health status by applications dedicated indices to aquatic vegetation based on specific endmembers collected in the field and wavelet analysis.

63 Mantua lake system, Water Adjusted Vegetation Index (WAVI) map derived from APEX data for September 2011 (left). Spectral response of different aquatic vegetation types and groups derived from APEX (right). (Villa et al., 2014). [Credit: CNR-IREA]

64 MULTITEMPORAL ASSESSMENT OF MACROPHYTES USING AQUATIC VEGETATION INDICES [Credit: CNR-IREA] Multispectral peak of season Multitemporal WAVI series

65 WP 5.8: Primary production Leader: USTIRLING Rationale C-fixation by phytoplankton is a key contributor to lake ecosystem energetics Tightly coupled to meteorology, climate and the catchment Model developed for ocean waters, but not tested in lakes Objective To develop a prototype model for the estimation of phytoplankton primary production in lakes from EO data.

66 Empirical VGPM Wavelength resolved Tilstone et al. (2009) Deep Sea Res. 56:

67 WP 5.9: Chlorophyll fluorescence Leader: U STIRLING Rationale Estimation of chlorophyll is problematic in lakes at low concentrations, especially in presence of high CDOM Fluorescence signal at 681nm might provide more accurate chlorophyll estimates Variability in relationship with chlorophyll related to physiology (photocompensation) Objective To undertake an evaluation of algorithms for the retrieval of chlorophyll fluorescence and concentration in lakes and explore relations with phytoplankton physiology

68 Chlorophyll-a in the Great Lakes derived from SICF peak at 685 nm. [Gons et al 2008]

69 WP6: EO-model integration (Leader: Deltares)

70 Delft3D: tool for effect chain analyses Physical parameters Transports (SPM,..) Water quality Ecology Fish, Birds other user functions, etc.

71 General modelling approach Delft3D-FLOW Hydrodynamics Delft3D-SED Suspended particulate matter (SPM) Delft3D-WAQ Origin of water and residence time Delft3D-ECO (BLOOM) Nutrients and primary production model

72 BLOOM BLOOM is a multi-species phytoplankton model Competition between phytoplankton types is the guiding principle in BLOOM BLOOM selects the optimum composition based on the ratio of the net growth rate and the requirements for each environmental resource Trade-off principle between growth and requirement: Relatively high potential growth rates may compensate a relatively large requirement hence opportunistic species win when light is high, efficient species win when there is little light