Global Earth Observation for integrated water resource assessment
2629 Hv Delft
Higher or Secondary Education Establishments
€ 1 021 688,50
Jaap Schellekens (Dr.)
Sort by EU Contribution
EUROPEAN CENTRE FOR MEDIUM-RANGE WEATHER FORECASTS
€ 504 872
€ 609 944,75
€ 412 236
CONSIGLIO NAZIONALE DELLE RICERCHE
€ 452 623
KENTRO KAINOTOMON TECHNOLOGION AE
€ 413 120
UNITED NATIONS EDUCATIONAL, SCIENTIFIC AND CULTURAL ORGANIZATION -UNESCO
€ 204 270,54
NATURAL ENVIRONMENT RESEARCH COUNCIL
€ 372 663
JRC -JOINT RESEARCH CENTRE- EUROPEAN COMMISSION
€ 259 350
CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE CNRS
€ 290 724
TECHNISCHE UNIVERSITAET WIEN
€ 336 284
PML APPLICATIONS LTD
€ 378 293,21
€ 316 600
UNIVERSIDAD DE SANTIAGO DE COMPOSTELA
€ 384 986
€ 311 160
UNIVERSITA DEGLI STUDI DI ROMA LA SAPIENZA
€ 328 696
€ 396 651,84
AMBIOTEK COMMUNITY INTEREST COMPANY
€ 138 772
€ 159 933
ADDIS ABABA UNIVERSITY
€ 160 396
UNIVERSIDAD NACIONAL DE COLOMBIA
€ 160 700
€ 159 400
SEVEN ENGINEERING CONSULTANTS OE
€ 279 500
INSTITUTE OF WATER MODELLING
€ 159 366
€ 141 400
OBSERVATORIO DEL EBRO FUNDACION
€ 159 389
INTERNATIONAL CENTRE FOR AGRICULTURAL RESEARCH IN THE DRY AREAS
€ 160 000
STICHTING IHE DELFT INSTITUTE FOR WATER EDUCATION
€ 196 768,16
Grant agreement ID: 603608
1 January 2014
31 December 2017
€ 11 344 199,01
€ 8 869 787
Final Report Summary - EARTH2OBSERVE (Global Earth Observation for integrated water resource assessment)
eartH2Observe integrates available global earth observations (EO), in-situ datasets and models to construct a global water resources re-analysis dataset of significant length (several decades). The resulting data allows for improved insights on the full extent of available water. New EO data sources have been tested, extended and existing processing algorithms have been combined with data from multiple satellite missions to improve the overall resolution and reliability. The usability and operational value of the developed data have been verified and demonstrated in a number of case-studies across the world allowing end-users to improve the efficiency of regional water distribution. All available data is disseminated through an open data Water Cycle Integrator portal to ensure increased availability of global water resources information on both regional and global scale. The data portal is part of the GEOSS data core water.
A multi-model ensemble-based global water resources reanalysis (WRR) existing of simulations from state-of-the-art land surface models and global hydrological models forced with the most-accurate global meteorological forcing based on re-analysis, satellite and in-situ data covering the period 1980 to 2014 with a spatial resolution of 0.25x0.25 degrees was created. This ensemble was carefully validated and showed the weak points and strengths of the generated dataset providing good guidance of the applicability in water resources management.
The value of the data for water resources assessments and created EO datasets was tested in eight countries covering five continents during end-user workshops. eartH2Observe datasets have been actively used in all case-study countries by the research institutes from these case-study countries working in close collaboration with the European partners.
The Water Cycle Integrator - the project’s data portal - contains the final releases of eartH2Observe EO and WRR datasets. The portal contains several user oriented features such as area of interest definition with geo polygons, additional plotting and analysis tools and collaboration tools that have already been successfully used in the case-study countries and for the education of students. The data is part of the GEO data core.
Project Context and Objectives:
eartH2Observe brings together the findings from European FP projects DEWFORA, GLOWASIS, WATCH, GEOWOW and others. It integrates available global earth observations (EO), in-situ datasets and models to construct a global water resources re-analysis dataset of significant length (several decades). The resulting data allows for improved insights on the full extent of available water and existing pressures on global water resources in all parts of the water cycle. The project will support efficient and globally consistent water management and decision making by providing comprehensive multi-scale (regional, continental and global) water resources observations. It will test new EO data sources, extend existing processing algorithms and combine data from multiple satellite missions in order to improve the overall resolution and reliability of EO data included in the re-analysis dataset. The usability and operational value of the developed data will be verified and demonstrated in a number of case-studies across the world that aim to improve the efficiency of regional water distribution. The case-studies will be conducted together with local end-users and stakeholders. Regions of interest cover multiple continents, a variety of hydrological, climatological and governance conditions and differ in degree of data richness (e.g. the Mediterranean and Baltic region, Ethiopia, Colombia, Australia, New Zealand and Bangladesh). The data will be disseminated through an open data Water Cycle Integrator portal to ensure increased availability of global water resources information on both regional and global scale. The data portal will be the European contributor to the existing GEOSS water cycle platforms and communities. Project results will be actively disseminated using a combination of traditional methods (workshops, papers, website and conferences) and novel methods such as E-learning courses and webinars that promote the use of the developed dataset.
The eartH2Objective overall objective has been divided and refined in 7 specific objectives to be pursued during the project: (1) Validate EO products based on end-user needs and metrics ensuring the value of the project’s final datasets for local and regional decision making. (2) Test new EO parameters and data sources in order to improve monitoring capabilities in terms of resolution and reliability and to explore techniques for synergetic use of datasets from multiple satellite missions. (3) Integrate in-situ data on groundwater, surface water, water quality, soil moisture, precipitation and evaporation with EO-driven models resulting in a model and multi-data global water resources reanalysis. (4) Assess error propagation through large scale water resource modelling, using in-situ data from data-rich validation sites to be able to provide the project’s datasets to end-users together with their associated error characteristics. (5) Develop a global reanalysis of water resources that supports efficient water management and decision making by boosting the availability of information on freshwater resources worldwide. (6) Demonstrate the usefulness of the integrated water resources time series at the operational level in regional and local case studies. (7) Ensure the results will become part of the GEOSS Data-CORE encyclopaedia, connecting to the GEOSS Water Cycle Integrator (WCI) initiatives.
In WP2 the value of the global EO and modelled datasets was evaluated in the support of the implementation of European Policies, river basin planning and for the derivation of global water scarcity maps. The data requirements of the different policies were investigated and gaps were identified. With multi-criteria score cards the value of the datasets in support of processes at different levels of policy making was assessed. From the analysis it can be concluded that selected variables (i.e. snow, soil moisture and evaporation) have significant potential because observations are limited. Furthermore, the spatial consistency of the datasets provides significant advantages for European wide assessments and trans-boundary basins where in-situ data availability may be asymmetrical across the borders of the riparian countries.
In WP3 EO-derived data products and time series were constructed for different components of the terrestrial water cycle. The WP tested new data sources, improved existing algorithms and created a long-term database of Earth Observation products existing of Precipitation (incorporating a.o. TMPA-3B42, CMORPH, GSMaP_MVK, GSMaP_Gauge_RNL, PERSIANN, TAMSAT, and RFE2), Soil moisture (ESA CCI Soil Moisture product), Evapotranspiration (GLEAM), Inundation and flood extent (based on LandSat8, MODIS and SAR), Lake water level (radar altimetry), Snow cover (MODIS), Water quality (> 20 satellites including Sentinel 1 and 2). Furthermore, soil moisture imagery was used to reduce the uncertainty in rainfall estimation. A global 0.1° resolution Multi-Source Weighted Ensemble Precipitation (MSWEP) was developed. Algorithms for the detection of snowfall were improved. Finally, use of satellite X-Band Synthetic Aperture Radar data for the detection of precipitation areas has contributed to the design of a new SAR mission (named KydroSAT) in the H2020 program.
In WP4 we have implemented and further developed the international land model benchmarking system ILAMB (https://www.ilamb.org/). A number of ydrological benchmarking datasets have been added to the evaluation package that were developed as part of the eartH2Observe project (ESA-CCI soil moisture, GLEAM-v3 evapotranspiration, MSWEP precipitation). ILAMB already included MODIS snow and GRACE terrestrial water storage data. Using these novel datasets and others we have evaluated WRR1 and WRR2 as well as the ensemble (WRR-ENS) simulations that included the assessment of the influence of precipitation uncertainty by using a set of satellite precipitation products.
In WP5 a multi-model ensemble-based global water resources reanalysis (WRR) consisting of simulations from state-of-the-art land surface models and global hydrological models forced with the most-accurate global meteorological forcing based on re-analysis, satellite and in-situ data covering the period 1980 to 2014 with a spatial resolution of 0.25x0.25 degrees was created. It includes a.o. soil moisture, precipitation, discharge, snow, evaporation, groundwater. An additional model ensemble was run with four different precipitation datasets to assess the uncertainty in the ensemble product. The performance of the WRR was evaluated using the iLAMB benchmarking system with ESA-CCI soil moisture, GLEAM-v3 evapotranspiration, MSWEP precipitation, GRACE and MODIS data.
In WP6 the value of the data for water resources assessments was tested in eight countries covering five continents. WP6 also provided feedback to WP3 (EO datasets) and WP5 (water resources reanalysis) to apply and improve the eartH2Observe data products for water resources analysis and water management at the river basin level through involvement of important end-users. The needs of the end-users were assessed for each case-study in the first year in local workshops. The results of the case studies were disseminated during a second round of end-user workshops in close collaboration with WP8 in the last year. The research institutes from the case-study countries worked in close collaboration with the European partners and after the first year the uptake of the eartH2Observe datasets in the case-studies highly improved. For each case study a specific added value could be found.
The primary role of WP7 was the provision of a central repository to host the data produced in the other work packages, and the provision of a suite of online tools for data access, analysis and sharing. This has led to the development of the Water Cycle Integrator (wci.earth2observe.eu). The portal hosts data from WP3 and WP5. A number of tools have been developed with non-technical users in mind, ensuring that there are few barriers to accessing, viewing, spatial and statistical analysing or downloading data. In the last year the project’s server has been connected to the GEOSS data portal and all datasets can now be viewed here as well. Finally, an additional snow portal has been created that was designed in close interaction with the EEA.
In WP8 we have disseminated the project outcomes through a large number of dissemination activities. 110 scientific papers have been published. A set of eLearning courses is now available. One of these was prepared as part of a conference and related training (WaterNet) with African students and stakeholders in Namibia. In the last year the project team hosted sessions dedicated to eartH2Observe at the EGU conference in Vienna and the European Geo Workshop in Helsinki. Three special issues were / will be published as part of eartH2Observe in Hydrology and EartH System Sciences, Water Resources Management and in International Journal of Applied Earth Observation and Geoinformation.
This section describes the state-of-the-art at the project’s outset, the outlined ways in which the project planned to progress beyond the-state-of-the-art and the project’s achieved advances. The information is organized according to the project’s Work Packages and tasks.
WP2 Policy support
The communication issued by the EU in 2012, reviewing the implementation of water resources policies by member states, “A Blueprint to Safeguard Europe’s Water Resources, underlined the challenge to managing Europe’s water resources sustainably. Trends demonstrated a wider spread and increased frequency of water scarcity and stress, reinforcing the need for a more sustainable management of water resources in Europe. The blueprint identified the need of an improved knowledge base and tools to support member states in implementing the policies such as the Communication on Water Scarcity and Drought, the Floods Directive, and the Water Framework Directive. The issue of increasing water scarcity and stress is, however, not limited to Europe, with the growing threat to water security at the global level. Within the policies that have been developed in Europe, the use of indicators as metrics to support decision making towards improved sustainability in the management of water resources is a key concept. Establishing indicators, such as for example an indicator that defines water use as being unsustainable when the total abstractions exceed the renewable water resource, requires consistent datasets of sufficient spatial-temporal resolution and with a sufficient period of record. Globally, the availability of consistent and accessible datasets is equally identified as a challenge in establishing meaningful metrics to support decision making in support of water security, not least related to the Water-Energy-Food nexus in transboundary basins. The development of earth observation and water resources reanalysis datasets, such as undertaken by EartH2Observe, offers significant potential in providing datasets that can be used to fulfil these policies, or in establishing meaningful indicators of the occurrence of drought and water scarcity at regional and global level. Additionally, these datasets may offer value to decision making at the local level, particularly earth observation datasets; augmenting datasets from in-situ data networks either by providing additional parameters where such networks are well developed such as in many countries in Europe; or where in-situ data networks are poorly developed or difficult to access.
WP2 Planned progress beyond the state of the art
The main question that Work Package 2 (WP2) set out to progress on, is the question of how the comprehensive datasets and information developed within the E2O project; in particular within the work packages 3, 4 and 5; can support and provide added value to decision making in managing water resources; and how these datasets can support the implementation of water related policies. In answering this question, and the specific objectives posed; the work package was divided into four tasks. Developing this analysis was closely related to other work packages, not just those that focused on datasets themselves (Work packages 3, 4 and 5), but also the case studies work package (Work Package 6) as through these case studies collaboration could be sought with end-users. Work Package 7, which developed the data portal that makes the datasets developed in the project accessible to end users, was also important, as it is through this portal that the potential of the datasets can be fulfilled.
The first two of the four tasks identified in WP2; Task 2.1 (Comprehensive datasets and information in decision making and policy – EU perspective) and Task 2.2 (Comprehensive datasets and information in decision making -- GEOSS Workbenches and International River Basin Perspective); set out to reveal the potential contribution in the implementation of policies the datasets have in complementing existing datasets, or in filling gaps where existing data sets are lacking,. The first of these two tasks focussed on selected water related European policies, while the second focused on selected policies within the international basin case studies. Task 2.2 also set out to explore links to the global water policy agenda, in particular to GEOSS.
The objective of Task 2.3 (Comprehensive datasets and information in decision making – Global Perspectives) was to explore the contribution of the datasets in providing consistent information to support monitoring of drought and water scarcity at the global scale. The objective of Task 2.4 (Value of comprehensive datasets and information– user oriented metrics and operational perspectives) was to develop and apply frameworks through which the value of the datasets in supporting operational decision processes in managing water resources, such as in planning and operation of irrigation districts, can be assessed, and how these decisions processes can potentially be augmented through the datasets by reducing uncertainty.
WP2 Project advances
(1) Value of Comprehensive datasets and information in decision making and policy – EU perspective
To assess the value of the datasets such as those developed in the project to the implementation of European polices, an end-user oriented framework was developed that comprised of three steps. First a detailed assessment of data needs in the policies was developed. This is followed by a gap analysis, and in the third step a multi-criteria score card approach is developed (see the example in Table 1). This assesses the suitability the datasets considering on the one hand intrinsic properties such as spatial and temporal resolution, period of record and ease of use, and considerations such as the complementarity and availability compared to existing datasets. Based on these assessments the key messages on the value of the datasets developed in the EartH2Observe project include:
- The spatial and temporal scales of the datasets is generally coarser than currently available high quality national datasets, or pan-European datasets and in this respect provide little added value. Further improving the resolution of these global datasets, or the development of down-scaled local re-analysis datasets could improve these deficiencies.
- Selected variables in both the EO datasets and the variables in the WRR are identified to have significant potential. These include all variables for monitoring snow (extent of snow cover as well as water equivalent), evaporation and soil moisture. Remotely sensed variables of water quality also provide significant potential. This potential is found primarily due to the low density of in-situ data for these variables across Europe, or due to the significant cost reduction over in-situ techniques.
- The spatial consistency of the EO and WRR data provides significant advantage where comparison across countries is required (such as in the European Environment Agency Water Accounting framework), or in trans-boundary basins where in-situ data availability may be asymmetrical across the borders of the riparian countries.
(2) Comprehensive datasets and information in decision making -- GEOSS Workbenches and International River Basin Perspective
The same framework that was applied in assessing the value of the datasets developed in the project to water related policies in the European perspective, was applied to selected datasets in international river basins, focussing on the river basins in which cases studies were developed. Within these, the availability and accessibility of existing datasets was found to be quite a bit more varied than in the European perspective. In Ethiopia for example, a comparatively data poor country, issues with data availability and access mean that consistent and open access datasets such as provided by the global datasets have significant opportunity to complement the scarce existing datasets. In the case of countries where the availability of data is better developed, such as Colombia this is, however, somewhat different. At national level the opportunities the datasets have are recognised, particularly given the advantage that these are consistent across the country, with a sufficient period of record for establishing reliable national water balance studies. At the regional level, reflecting the case of Europe, the added value in complementing existing datasets was found to be quite a bit more limited. An important finding is the potential these globally consistent datasets provide to complementing national datasets in trans-boundary rivers basins. In many parts of the world, particularly in Africa, major river basins are transboundary, while the data availability in different riparian countries may be very much asymmetrical. This suggest that a key contributions of water resources re-analysis datasets is in addressing the Water-Energy-Food nexus in transboundary basins. Integrated decision support tools, such as WaterWorld, which was applied in several of the case studies in Earth2Observe were found to be important in realising the potential of the datasets through making these more accessible to decision makers and link these to the decision processes they deal with.
An assessment of the opportunities the datasets developed in EarH2Observe to monitoring the targets of the Sustainable Development Goals (SDG’s), which was identified as an increasingly important objective within the GEO programme and GEOSS, revealed these to be substantial in 75% of the goals and their underlying targets. Clearly these opportunities are most relevant to Goal no. 6, which is the main water-related Goal, but it was found that 75% of goals include targets that are directly or indirectly related to water.
(3) Comprehensive datasets and information in decision making – Global Perspectives
The data needs from the global perspective are quite different, and less related to the fulfilment of particular policies. The focus of our work was on the data needs of global scale users in global-scale drought and water scarcity studies such as the Red Cross/Red Crescent, the Food and Agriculture Organisation (FAO), and the World Bank. The water resources datasets developed provided the means to develop several drought and water scarcity metrics to assess severity, extent, frequency, duration and exposure to drought and water scarcity at the global and regional scale. Given the ensemble of models included in the water resources reanalysis developed in EarH2Observe this assessment revealed the current uncertainty (or model spread) in estimated water resources availability, especially in the relatively dry regions. However, it was found that in the case of the drought indicators these are less vulnerable to the model uncertainty, as these indicators represent anomalies with respect to their own climatology. In contrast, the modelling uncertainty in estimated water resources availability does, however, play a crucial role in absolute estimates of water scarcity.
The reliability of the estimates of global and regional freshwater resources availability was found to improve significantly when moving from the lower resolution Tier1, to the higher resolution Tier2 datasets. However, it was found that the improved resolution did little to improve the ability of the models to reflect the temporal variability of the indicators as the correlation of models did not improve significantly at the global scale.
(4) Value of comprehensive datasets and information– user oriented metrics and operational perspectives
The framework assessing the value of datasets developed within the EartH2Observe project to supporting local water resources decision processes, takes a user oriented approach. In selected cases an actual economic value was assigned to the datasets, based on a hydro-economic evaluation. The focus is on four local decision processes; as examples of the vast range of decision processes within the context of managing water resources. The four decision processes selected include (i) identification of flood events in the Limpopo basin in Southern Africa, in support of predicting flood impacts by flood managers in areas with little data; (ii) estimation of snowmelt volumes and snowmelt freshets in the Canadian Arctic to support operation of hydropower reservoirs; (iii) estimation of water resources availability to support the planning of command area size of irrigation districts; and (iv) supporting operational decisions by farmers on the planting of crops, subject to availability of water. In all four decision processes the results show the datasets indeed potentially add value through informing decisions. However, that value depends very much on the decisions; how data is used to inform the decision; and the impact of uncertainty that is inherent to these datasets and how this uncertainty influences the outcome of the process that the decision is made on (for example in the balance between investments in utilising water resources, and ratio to the return on those investments). From the four decision processes considered some key messages emerge:
• The resolution of the datasets may limit the scale to which these are applicable to local decision processes. In identifying flood events in Southern Africa, the lower resolution 0.5 degrees data sets provide skill identifying events in catchments of 2500 km2 or larger, while comparable performance was already found at 520 km2 for the higher resolution 0.25 degrees datasets. The improved resolution of in particular the MSWEP forcing dataset also added significant value to the simulation of snow accumulation and ablation in the Canadian Arctic. Further improvements to resolution could add further value and local relevance of global datasets.
• Despite the improvements found through increased resolution, decisions informed using the anomalies in the water resources datasets are more reliable than when using the absolute values, as the models may still display significant bias, even the higher resolution Tier 2 datasets. The use of (multi model) ensembles may also lead to more robust decisions being made, also when using absolute values.
• Hydro-economic frameworks are useful tools to model the decision process, and how this can be informed by the datasets. The added value of data and information can be also be made explicit. These also can help identify when hydro-meteorological uncertainty that is inherent to the datasets is important to the outcome of the decision process, and when it is of lesser importance. It may be that the hydro-meteorological conditions under which the uncertainty in the water resources estimation is greatest are not those that have the greatest influence on the outcome of the decision process. This can be made explicit through hydro-economic frameworks that make explicit where these datasets have value in supporting decisions, as well as how improvements to the models and data can increase that value.
WP3 Earth Observations – Combining and improving EO processing techniques
A thorough integrated assessment of the Earth's water resources requires the availability of suitable global products on all components of the terrestrial water cycle. Recent studies in hydrology, Earth observations, Earth system modelling and climatology have demonstrated that no single data source exists either from the ground or from space that is capable of answering the needs of a large variety of end users. Thus, testing new EO data sources, extending existing processing algorithms, and combining data from multiple satellite missions in order to improve the overall resolution and reliability of EO data included in the re-analysis dataset is the basis for a successful strategy in water resources assessment. WP3 had the major objective of constructing EO-derived data products and time series showing optimal performance for all components of the terrestrial water cycle. It aimed at testing existing and new data sources, innovative statistics and algorithms to improve the resolution and the reliability of monitoring techniques of regional and global water resources.
WP3 Planned progress beyond the state of the art
The progress beyond the state of the art regarded the following water cycle components: precipitation, soil moisture, snow cover, evaporation, inundated area extent, lake water level and water quality assessment. The planning of the progress regarded two main topics: 1) provided enhanced space time resolution and coverage datasets, and 2) improve performances of the retrieval algorithms (through the enhancement of their detection performances and the attainment of better validation figures). The following algorithms were planned for substantial improvement throughout the project:
• Precipitation: CNR-ISAC (CDRD, PNPR, 183-WSL), Univ. Roma La Sapienza (X-band SAR), CNR-IRPI (SM2RAIN)
• Soil moisture: ESA CCI (TU Wien)
• Evaporation: GLEAM (VUA)
• Inundated areas: GIEMS (Estellus)
• Surface water extents (Deltares)
• Convection and overshooting top detection (CNRS)
• Snow cover: MODIS and MSG-based algorithms (GISAT)
• Aquifer characterization (USC, Deltares)
• Lake water quality focused on lakes in Estonia (PML)
Improvements were sought especially in the direction of:
• Attain better detection skills (low intensity and snowfall detection from passive and active microwave remote sensing; use of combined satellite and model data to improve rainfall detection over complex terrain; detection of inundation in problematic areas such as those covered by forests and coastal areas);
• Devise higher precision products (lake water level using novel radar altimetry approaches; better description of evaporation through update the water balance model, better description of infiltration rates, and update the evaporative stress functions; lake water quality using 300 m ESA Envisat MERIS data and a water class approach),
• Provide better and more continuous space-time coverage (e.g. snow cover, soil moisture, precipitation).
• Introduce new products for an enhanced global coverage. Above all, soil moisture, precipitation, evaporation and inundated areas.
WP3 Project advances
WP3 has established a long-term database of Earth Observation products: Precipitation (17 products), Soil moisture (1), ET (2), Inundation and flood extent (2), Lake water level (3), Snow cover (3), Water quality (1). Each dataset includes advances over the datasets available at the start of the project. The advances are discussed below per variable which refer to the sub-task.
For precipitation the following improvements have been introduced:
• Production of a comprehensive archive of satellite precipitation products from existing state-of-the-art algorithms (TMPA-3B42, CMORPH, GSMaP_MVK, GSMaP_Gauge_RNL, PERSIANN, TAMSAT, and RFE2) and newly-developed CNR algorithms (PNPR, PNPR2, CDRD, and 183-WLS) in netCDF format (CNR).
• Development of all-sky frozen surface classification scheme for ATMS and analysis of GMI high frequency channels sensitivity to snowfall to be implemented in snowfall detection and retrieval schemes (CNR).
• Development of a 3-hourly 0.25° degrees global gridded precipitation dataset for the period 1979-2014. The Multi-Source Weighted Ensemble Precipitation (MSWEP). This dataset is based on averaging precipitation anomalies from the different datasets with bias correction using streamflow from about 15,000 catchments around the world. MSWEP was recently evaluated against other precipitation datasets using gauge observations and hydrological modelling (Beck et al., 2017b).
• Development and validation of a new processing framework in which satellite X-Band Synthetic Aperture Radar data are exploited to detect precipitation areas and quantify their intensity (Univ. Roma La Sapienza). The design of a new SAR mission (named KydroSAT) in the H2020 program, specifically devoted to precipitation and hydrological applications stemmed from eartH2Observe activities.
• Improved datasets of multi-satellite precipitation and convection (DC) daily occurrence from passive microwave sensors over the Mediterranean Basin and the Tropics (CNRS). The spatial resolution is 0.20° lat-lon, and temporal coverage 1999-2016 (still continuing). Improvements included an ameliorated error characterisation, the possibility of detecting overshooting convection –COV, and the characterisation of cloud microphysics using co-located CloudSat/Calipso observations. Their potential for evaluating regional models was demonstrated.
• KKT-ITC has presented a methodology towards a hybrid precipitation product that dwells upon the spatiotemporal accuracy of satellite precipitation estimates and enhances the magnitude underestimation that is often evident over mountainous and complex terrain areas. Considering the lack of in-situ observations in such areas, the NWP-adjusted satellite estimates can provide a viable substitute wherever the availability of gauge or radar rainfall data is not sufficient to drive the correction process.
TU Wien has considerably expanded and improved its multi-satellite soil moisture product from combining 11 active and passive microwave sensors (ESA CCI Soil Moisture product). The dataset features a thoroughly validated long time series (1978-2016) with extensive metadata (uncertainty, flags, etc.). The following improvements were achieved: a) integration of new sensors: SMOS, MetOp-B; b) improved accuracy/reduced uncertainty; c) simultaneous use of up to 5 sensors; d) reduction of the data gaps; e) implementation of new uncertainty estimation and propagation scheme to provide per-observation uncertainty estimates. Moreover, the multi-satellite dataset has contributed to a reduction of rainfall estimation uncertainty (ITC).
The ESA CCI Soil Moisture product is in preparation for operational services under the Copernicus Climate Change Service (C3S).
By combining information on precipitation with soil moisture data the uncertainty in precipitation estimates could be improved. Hereto CNR (IRPI, in collaboration with ISAC) and TU Wien have developed a new precipitation estimation method - SM2RAIN - where soil is assumed to act as a natural rain gauge for measuring the amount of rainfall fallen on the ground. The inversion of the soil water balance equation is done for retrieving rainfall from soil moisture data. Rainfall data (mm/day) have daily temporal resolution, 0.25° spatial resolution and are available for the period 1998-2015.
The GLEAM (Global Land-surface Evaporation: the Amsterdam Methodology) dataset was updated and temporally extended by VUA. Three are the main improvements in the new version of the algorithm: 1) a new data assimilation scheme has been developed, implemented and evaluated, evaluated, and assimilates the multi-sensor ESA CCI Soil Moisture product that was co-developed in the project (Martens et al., 2017, 2016); 2) the water balance module has been updated, describing infiltration rates as a function of the vertical gradient in soil moisture (Martens et al., 2017); 3) the evaporative stress functions have been updated based on experimental evidence (Martens et al., 2017).
Estellus has improved the Global Inundation Extent from Multi-Satellite (GIEMS) dataset, in particular for coastal areas that are very important because most of the global population lives in coastal cities. Various downscaling approaches were developed depending on auxiliary information in the VIS/IR (MODIS, Aires et al., 2014), SAR (Aires et al., 2013), or a Digital Elevation Model (DEM). A floodability index was built, which was then used to downscale the coarse-resolution dataset towards the high-resolution (90 m) GIEMS-D3 (Aires et al., 2017). GIEMS-D3 was evaluated over various the Amazon, Ganges-Brahmaputra, Niger, Canada, and Vietnam showing very good results (Aires et al., 2018). It is shown that VIS-based datasets cannot detect water below vegetation and are sensitive to cloud presence too. Therefore, Landsat-based datasets need to be combined with GIEMS-D3 to fully describe all type of wetlands.
A novel method to quickly and reliably detect surface water and surface water changes was developed by Deltares using the computing power of Google Earth Engine to process large amounts of Landsat data. The result was demonstrated in the Australian Case and has been further picked up by the SERVIR project (https://servir.adpc.net/tools/surface-water-mapping-tool).
I-MAGE has produced a number of advances in inland water level retrieval using radar altimetry, there are: 1) development of new algorithms for data processing; 2) improvement of the data accuracy (few cm) thanks to a better understanding and analysis of the signal; 3) production of new or extended time series over lakes and small reservoirs at regional scale; 4) processing and tests for major rivers (US, Europe, Africa, Australia).
A new global water table depth high resolution dataset has been produced in WP3 improving on the current state-of-the-art product in Fan et al. (2013). The main advances are, first, that it is a dynamic, instead of steady-state calculation, thus providing information on groundwater's seasonal cycle, and, by covering the period from 2004 to 2014, suitable for comparisons with GRACE (this validation is part of WP4). Second, the water table is fully coupled to the soil-vegetation above, discretized in 40 layers that provide high resolution in the top 20 m. Infiltration is computed solving the Richard's equation and root uptake is dynamic, driven by an energy efficiency function. Third, rivers and lakes are represented and interact via two-way fluxes with the soil and the groundwater. Floodplain and inundation processed are accounted for. And finally, the model is driven by the high quality global meteorological forcing from E2O WRR2. The spatial resolution is kept at 30" globally, but all the aforementioned new developments yield a much more realistic product than the previous version and result in a unique dataset providing a global view of water table depth patterns.
Data from MODIS sensor (Aqua and Terra satellites – MOD10A1 and A2, MYD10A1 and MYD10A2) has been downloaded, processed and mosaicked by GISAT for the period 2000-2016. Daily data (MOD10A1 and MYD10A1) provides information about Snow cover, Snow spatial QA, Snow albedo and Fractional snow cover. 8-day composites (MOD10A2 and MYD10A2) provide information about Maximum snow extent and 8-day snow cover. Derived local data has been processed based on different research needs during the project duration.
PML carried out comparisons of existing algorithms for Chlorophyll-a and coloured dissolved organic matter (CDOM) in support of the WP6 Case study in Estonia. During eartH2Observe, the validation of new global algorithms was conducted over Lake Peipsi. The time series of annual 90th percentile Chlorophyll-a (2002-2012) showed areas most affected by blooms in Lake Peipsi; the optical water typology showed the seasonal occurrence of e.g. cyanobacteria blooms. Finally, a comparison was made with watershed/nutrient modelling in the Estonia case study (Fink et al., in prep).
WP4 Validation of EO products and models using in-situ data
Error characterisation methodologies are needed to deliver meaningful EO products to users, providing not only actual values, but also estimates of their accuracy. Moreover, knowledge of the relative errors for each data source allows for the development of adaptive combination algorithms required, as not all data sources / missions are available continuously. On longer time scales, error information will help to discriminate between error and actual natural variability of EO products. The International Precipitation Working Group has initiated an assessment of six global satellite rainfall products over continental regions: the Pilot Evaluation of High Resolution Precipitation Products (PEHRPP; Arkin et al., 2017)). Recent satellite rainfall error studies (Anagnostou et al., 2010; Stampoulis and Anagnostou, 2012) have shown that the accuracy of satellite rainfall data is subject to seasonal, storm type, terrain and climatological factors. Other recent studies demonstrated possibilities for the generation of ensemble satellite-rainfall products, in which each element is consistent with the original satellite data (Hossain and Anagnostou, 2006). Feeding these elements into hydrological models enables the ensemble-based characterisation of precipitation uncertainty propagation in runoff and other hydrological variables. Several novel error modelling techniques can be useful for EO validation, such as triple collocation and R-metrics. Here, an assessment with independent reference data and a range of statistical approaches is required to assess the retrieval errors. Most error characterisation methods rely on the use of reliable in-situ data (Brocca et al., 2011; Albergel et al., 2012). Therefore, GEWEX, CEOS, and GEO jointly promoted the initialisation of the International Soil Moisture Network (ISMN), which brings together and harmonizes SM datasets from several regionally operating networks worldwide. Alternatively, model-based SM datasets and satellite-based products of other variables (e.g. precipitation) are increasingly being used to assess the quality of the satellite-based SM retrievals (Dorigo et al., 2010) as the effect of assimilating remotely sensed soil moisture in land surface models on modelling skills is increasingly being regarded as a valuable quality measure (Bolten et al., 2010).
WP4 Planned progress beyond the state of the art
Error characteristics of the global raw and improved EO data and the global reanalysis dataset will be estimated in data rich sites using in-situ data and other EO data products as reference. The error characterisation will be based on the end-user oriented error metrics. The obtained error-characteristics will be interpolated over time and extrapolated to data sparse regions through error modelling. Based on the error-characteristics; 1) EO retrievals will be improved for use in hydrological decision making and 2) a selection of the locally optimal datasets can be made. The error characteristics of the overall global water resources reanalysis will be assessed by propagating the error characteristics of individual input datasets through the global modelling chain.
WP4 Project Advances
In order to highlight and evaluate changes/improvements as a result of ensemble model developments and use of EO data in the eartH2Observe suite of hydrological/land surface models from Tier-1 (WRR1) to Tier-2 (WRR2) results, we have implemented and further developed an international land model benchmarking system ILAMB (https://www.ilamb.org/). In particular, we have added hydrological benchmarking datasets to the evaluation package developed as part of the eartH2Observe project (ESA-CCI soil moisture, GLEAM-v3 evapotranspiration, MSWEP precipitation). Using these novel datasets and others we have evaluated WRR1 and WRR2 (Martinez-de la Torre et al., 2018), finding an overall improvement in WRR2 integrations, attributable in part to the higher resolution and accuracy of the new precipitation dataset MSWEP and in part to the model development carried out by the modelling groups using different EO datasets. Additionally, to the overall improvements introduced by WRR2, an analysis was presented in WP4/WP5 collaboration with the ILAMB evaluation of hydrologic simulations for the WRR-ENSEMBLE multi-model/multi-precipitation-forcing experiment.
At CEH a dry-down metric that can be used to evaluate a landscape and its response to atmospheric forcing independently of the accuracy of the precipitation input was developed. This is important as it can be used to evaluate models used in coupled land-atmosphere mode when the forcing is often not commensurate with the actual atmosphere. Our metric uses the most direct observation of drying: the rate of change of evapotranspiration after a rainfall event using eddy-correlation observations (commonly referred to as flux tower data). We analyzed the observed data to show how the drydown timescale is characteristic of different land cover types and this was then used to evaluate the WRR1 suite of global hydrological and land surface models. We showed that the data suggests that trees take longer to dry the soil than grasses and that land surface models capture this characteristic time-scale difference better than large-scale hydrological models. We therefore concluded that the drydown metric has value in understanding land-atmosphere interactions and model evaluation.
We have also been carrying out a study on the relative dominance of land surface model uncertainty and uncertainty in precipitation product used to drive the land surface models. The motivation of this is practical: in situations where model uncertainty is significant, then stakeholders using Earth2Observe results must be made aware of, and take account of, the range of predictions possible from standard model simulations. Alternatively, if precipitation product uncertainty dominates then users and stakeholders need to concentrate more resources on selecting the most appropriate product to use and to interrogate more strongly the potentially sparse data base of precipitation measuring stations on which the precipitation products are based. The significance of this study is that we will provide crucial practical advice on the reliability of Earth2Observe results and predictions in terms of extreme drought and flood events in each of the global study regions.
IVM-VU has evaluated the influence of different precipitation products in various global land and hydrological models for the successful identification and characterization of hydrological droughts at the global scale. Whilst most validation studies evaluate the ability to correctly assess hydrological means and extremes, we focused on the end-user. In doing so, we determined if significant changes occur in the ability to correctly identify and characterize hydrological drought conditions as a result of varying precipitation products. Moreover, we evaluated how the variation in results due to varying precipitation products compares to the variations in results due to varying hydrological or land surface models. The significance of this study is that we try to provide practical advice on where improvements in model or forcing data matter when it comes to drought policy making and/or where model or input forcing data improvement are of less added value.
Météo-France has customized snow variables derived from in situ observation data sets: HSDSD Version 2 snow depth and FSUHSS snow density, snow depth and SWE observations in Former Soviet Union available from the NSIDC data base. GRACE satellite data where used to verify the consistency between the model-simulated changes in the terrestrial water storage and the gravimetric observations. The snow depth and SWE historical in situ observations were used to validate the WP5 model simulations of snow cover. Global hydrological variables were assessed, before and after the assimilation of EO products in the models operated in WP5.
ITC-KKT has developed a novel algorithm for blending precipitation information from different data sources (satellite, reanalysis, etc.) with the aim to develop an improved precipitation product for hydrologic applications. As has been recently demonstrated for the Iberian Peninsula, the relevant improvements in precipitation estimates have a great impact on the simulation of hydrologic variables (e.g. streamflow). In turn, this can have a significant socioeconomic impact considering that improvements in streamflow simulations lead to improvements in flood prediction, reservoir operation, irrigation practices and water resources management in general.
In mountainous and complex terrain areas where in-situ observations are scant, a methodology of satellite adjustment that uses high-resolution NWP simulations provides a viable substitute in order to drive the correction process. In addition to the actual error correction, this allows for in-depth microphysical analysis of storms that are inadequately detected from the satellite instruments. The latter can provide valuable feedback to the retrieval community, restructure our perception of what is achievable in terms of satellite precipitation retrieval and contribute to shape the basis of future algorithm development.
WP5 Water Resource Reanalysis
The accurate mapping and estimation of global water resources requires the use of many sources of earth observations (such as satellite and ground-based remote sensing, in situ measurements, vertical profiles, etc.), combined with state-of-art earth system modelling components that are developed for hydro-meteorological and environmental applications. On global scale, there is a limited number of global reanalysis datasets that can support water resources. Among these, multi-model reanalysis of the state of the surface water storage and fluxes, like the GSWP2 or WATCH datasets, provide an ensemble that is not dependent on a single model. This is generally superior to the results of any individual. Considering the significant uncertainties in modelling the different components of the surface water cycle a multi-model approach can consider the inter-model uncertainties that can be used in downstream applications of water resources.
WP5 Planned progress beyond the state of the art
The aim of WP5 was to produce a multi-model ensemble-based global water resources reanalysis in which state-of-the-art land surface models and global hydrological models are forced by the most-accurate global meteorological forcing provided by atmospheric reanalysis including EO data. The time-series with a length of more than 30-years for the global water reanalysis (1979 to 2014) were generated for parameters of all compartments of the water cycle (soil moisture, precipitation, discharge, snow, evaporation) including water demand, water levels in rivers, lakes and reservoirs for a subset of modelling systems. This was be achieved by the following main develpments:
1. Generation of a first release (tier 1) reanalysis of the global water cycle.
2. Investigation of the value of earth observations to improve the representation of hydrological processes and enhance the quality of the land data assimilation systems.
3. Characterization of the uncertainties of the advanced reanalysis (or tier 2), which was build upon verification, data assimilation and model improvements.
4. Generation of an ensemble-based advanced global water reanalysis with associated error estimates.
All WP5 partners produced a first version water resources reanalysis (tier-1), adopting established methods and modelling platforms at the start of the project. This provided a benchmark reanalysis enabling the evaluation improvement within the project lifetime. The improvements included the use of a suite of precipitation products (e.g. higher resolution and error-adjusted from WP3 and WP4), which allowed the evaluation of the role of improved quality precipitation forcing in water resources assessment; also Land Data Assimilation Systems that ingest EO data, in interaction with WP3. Additionally, the WP5 partners developed improved land-surface and hydrological parameterizations for their modelling systems that are key for describing the water cycle evolution. These included refined models and model parameters, contributing to improved water estimates.
Based on these improvements, a second advanced version of the global WRR (tier-2) was produced and clustered in ensemble-based products characterized by a number of statististics: mean, median, standard deviation, 25th and 75th percentiles, minimum and maximum of the ensemble.
WP5 Project advances
The project managed to develop a multi-model ensemble-based water resources reanalysis dataset at a spatial resolution of 0.25x0.25 degree for the 1980-2014 period (WRR tier-2), based on eight state-of-the-art land-surface and hydrological models. The dataset uses the most accurate global meteorological forcing, including for the precipitation forcing the recently developed (within the project) Multi-Source Weighted-Ensemble Precipitation (MSWEP) dataset. Within eartH2Observe ECMWF developed a physically based method to provide the meteorological forcing for the WRR1 on 0.25 degrees instead of the 0.5 degrees resolution of WRR1. This increase in resolution was a request from all case-studies as well as those modelling groups that could run there global models at the higher resolution.
While moving from WRR1 to WRR2 the description of various physical processes in all the modelling systems that are part of the ensemble has been improved based on new EO datasets and assimilation thereof. The assimilation systems generally lead to an improvement in the simulation of the fields that are assimilated. The assimilation of river discharge allows also to estimate errors in evapotranspiration processes. The assimilation enables to use Earth Observations in near-real time applications. For nearly all models and most variables, the global water resources are better reproduced in the WRR tier-2 simulations compared to the WRR tier-1. The mean and median of the ensemble dataset generally outperform the single models for most of the considered variables.
All datasets are available on the project’s data server and from the individual model simulations the ensemble mean, min, max, std and 25, 50 and 75 percentile values have been calculated and uploaded.
A multi-model and multi-forcing ensemble-based water resources reanalysis dataset has been developed as well at a spatial resolution of 0.25x0.25 degree for the 2000-2013 period. Seven partners participated in the modelling exercise based on four different precipitation datasets (GSMaP, CMORPH, TRMM and MSWEP).
By using both multiple models and multiple forcing datasets the uncertainties associated to the precipitation forcing could be estimated taking into account the variation amongst models. The key surface fluxes and states simulated with the seven models have been analysed following the iLAMB approach (which will be discussed in the following paragraph). The analysis indicated that the difference in precipitation forcing increases the spread in these flues both at the global and basin-scale. The spread is larger than when compared with the multi-model ensemble forced by MSWEP. The mean and median of the multi-model multi-forcing ensemble dataset generally outperform the single models for most of the considered variables, confirming the value of the ensemble approach.
The WRR1, WRR2 and WRR-ENS model simulations have been evaluated using the International Land Model Benchmarking system (ILAMB). The system allows for a consistent inter-comparison of model results and established Earth Observation datasets. Simulated water fluxes and states have been evaluated against GLEAM V3, MODIS, FLUXNET-MTE, GLOBSNOW and GRACE as well as snow data from in-situ measurements (by ECMWF), LAI and surface soil moisture from satellite observations (by Météo-France) and river discharge observations (by CNRS).
WP6 Case studies/specific regional applications
This work package consisted of selected case studies to evaluate the applicability of the global hydrological model results and the new EO data and derived products for water management at river basin level. Moreover, this WP also provided feedback to WP3 (EO datasets) and WP5 (water resources reanalysis) to apply and improve the eartH2Observe data products for water resources analysis and water management at the river basin level through involvement of important end-users. The results of the case studies were disseminated during workshops in close collaboration with WP8 (dissemination). The WP6 objectives were:
- To consult the agencies responsible for water resources management for fine-tuning the user requirements with model types and EO products needed for efficient decision making, in close cooperation with WP2;
- to test the applicability of the improved EO data (WP3), the global water resources reanalysis outputs (WP5), and the Water Cycle Integrator (WP7) for water management purposes at the case study level;
- to combine the EO products (WP3) and global modelling outputs with river basin-scale hydrology models using the downscaling and nesting procedures developed in WP5 in order to improve assessment and prediction of water resources variability and availability, in close collaboration with the case study end-users.
WP6 Planned progress beyond the state of the art
The case studies were conducted in eight countries covering five continents. For each country a specific water management issue was identified and studied. It was intended to apply the products developed in WP3 and WP5 to improve the operational water management in the case study countries through involvement of important end-users. Six case studies were carried out in eight countries, which vary widely in available water resources:
1. The Mediterranean (Spain and Morocco) with a focus on drought, water resources availability and water management,
2. Eastern Europe (Estonia) with a focus on water quality,
3. Africa (Ethiopia) with a focus on extreme hydrological events and water resources variability,
4. South America (Colombia) focusing on flooding, water demands and water quality,
5. South Asia (Bangladesh) focusing on glacier-fed water systems and water availability,
6. Oceania (Australia and New Zealand) where the focus was on water resources, governance and the applicability of the EO products for the national assessment systems.
In these case studies it was intended to (i) apply or improve Earth Observation data and methods for specific hydrological purposes, and (ii) to enhance hydrological modelling capabilities in the eight case study countries. Especially in the developing countries (Bangladesh, Ethiopia, Colombia, Morocco) not much hydrological measurements and monitoring is done, which hampers the assessment and management of available water resources. The intention of WP6 was to assist those countries with better data and modelling capacities that will help to improve future water management. In the developed countries (Spain, Estonia, Australia, New Zealand) the intention was to test the eartH2Observe datasets and products by comparing those to in-situ measurements and model results. In those countries, the added value of the eartH2Observe products was evaluated. In particular, it was determined how accurate global Earth Observation and global hydrological modelling is for specific hydrological purposes.
WP6 Project advances
The focus of the work in Morocco was on drought and water management in the Oum Er Rbia basin. Two hydrological models (SWAT and PCR-GLOBWB) have been set-up for the basin, and runs were performed using in-situ and E2O forcing data. Model calibration was done using EO-based data of evapotranspiration and soil moisture. The models show a good capability to simulate river discharge patterns. In addition, the E2O WRR2 multi-model ensemble was used to analyse drought in the Oum Er Rbia basin and entire Morocco.
In Spain the focus of the research was the assessment of groundwater and drought. Two land surface models, LEAF-HYDRO and ISBA have been applied at high spatial resolutions (2.5 and 5.0 km) and forced by the E2O forcing (Tier 1 and Tier 2 forcing) data and a local product (SAFRAN). Groundwater recharge and groundwater levels were simulated for entire Spain. Using the E2O Tier 2 forcing gave better results than the previous Tier 1 forcing, due to the higher spatial resolution and improvement of the rainfall data. But in the mountainous regions where the still coarse resolution of the E2O forcing data leads to unrealistic rainfall values, the calculated recharge is biased.
In Estonia the research was mainly dealing with water quality of lake Peipsi. The case study combined hydrological modelling (WaterGAP3; WaterWorld), water quality modelling (WaterQUAL component in WaterGAP3; WaterWorld), EO analysis of water quality (MERIS data) and in-situ data. The results show that the two models are well capable of simulating total phosphorous loads (WaterQUAL), and water pollution by human impacts (WaterWorld) in lake Peipsi. The analysis of MERIS imagery provided accurate assessments of chlorophyll-a concentrations in lake Peipsi, with a high spatial and temporal resolution in the absence of a dense in-situ measurement network.
The Ethiopian case study focussed on the upper Blue Nile basin, which is one of the most important water sources in Africa. Water resources assessments in the area are hampered by a lack of reliable precipitation data. Global precipitation products were compared with in-situ data and tested for hydrological modelling of the upper Blue Nile. The testing of the different EO-based precipitation products showed the WRR2 forcing rainfall data performed better than all other EO-based datasets. The analysis was extended towards modelling using the CREST hydrological model, and this local model was compared with WRR1 and WRR2 global model results. The modelled discharges were compared with measured discharges at three stations. Overall the WRR2 models performed better than the WRR1 models. One of the global models (WaterGAP3) performed equally well as the local CREST model using in-situ data.
In Colombia, the case study was dealing with the water resources in the Magdalena river basin. Several models (DWB, VIC, wflow-hbv, WaterWorld) were set-up and tested with different forcing datasets. A first step in the research was the comparison of EO precipitation products with in-situ measurements. Rainfall data from the WRR2 forcing dataset performed best and was much improved compared with the earlier coarse resolution WRR1 forcing data. Also, a comparison between WRR2 global model discharge results and the measured values at many sub-basins was made. Some of the uncalibrated global models provided good estimates of discharge. Model simulations of local and global hydrological models could be improved by assimilating satellite observed river discharges. The hydrological modelling was used to assess water availability in the Magdalena basin using different indicators. Such information is useful for basin-level planning and water management.
The focus of the Bangladesh case study was on the quantification of water resources variability in the Brahmaputra river basin. Several EO-based rainfall products were compared with in-situ measured rainfall. On a monthly scale, the EO-based products performed well, but were much less accurate on a daily scale. The WRR2 forcing rainfall data showed the best correlation with the measured rainfall data. The MIKE-BASIN model was used to simulate river flows using in-situ and EO-based rainfall. All satellite-based products resulted in an underestimation of discharge, with TRMM rainfall giving the best results. The WRR1 and WRR2 global hydrological simulations were also compared with measured discharges. The global models generally underestimate the peak discharges, but the ensemble mean gave reasonable results and can be used for flood analysis in the Brahmaputra basin. As a final step, a bias correction to the ensemble mean was made which provides an even better representation of measured discharge, even better than the locally modelled discharge with MIKE-BASIN.
In Australia, EO-based soil moisture products were tested and used for hydrological model assimilation in the Murrumbidgee river basin. The original AMSR-E soil moisture and a new, sharpened AMSR-E soil moisture product were compared with measured and modelled soil moisture over entire Australia. Both products showed a good agreement with the measured and modelled soil moisture. The sharpened AMSR-E soil moisture product was used for data assimilation in the PCR-GLOBWB hydrological model for the Murrumbidgee basin. Using soil moisture in the assimilation resulted in more improvement in discharge modelling than using measured streamflow. Combining the two types of data in model assimilation gave the best improvement. In addition, five global models from the WRR2 were used to evaluate the amount of water available for irrigation. The developed methodology shows that the hydrological modelling provides a potential improvement in irrigation planning.
In New Zealand the aim of the case study was to improve assessment of groundwater levels and recharge rates at national and catchment scales using EO data. A rainfall-groundwater recharge model was developed to estimate groundwater recharge. The model uses inputs of rainfall and actual evapotranspiration (AET). The AET estimates were derived from the Modis MOD16 product. For groundwater levels, the equilibrium groundwater table (EWT) approach was applied. National-scale results for rainfall recharge, evapotranspiration, and groundwater table levels have been produced and are available for national water management planning. The results indicate that the MOD16 evapotranspiration product is a good interpolator between station-based estimates, but it risks large overestimations in steep sloping terrain of New Zealand.
WP7 Water Cycle Integrator & Open Data Access
The suite of tools available to users to view, explore and analyse geospatial data at the outset of the project was already considerable. However, these tools often required users to install software on their own machine, which could cause compatibility issues with the users’ operating system or other installed software. These tools are often disjointed requiring users to utilise multiple software tools to achieve their desired results, and in many cases the user is required to download the source data before being able to explore and analyse them. This requirement to download data, which can often extend to many gigabytes or, in some cases, to the terabyte scale, is barrier that many users cannot surmount.
Software developed by Unidata called THREDDS Data Server offers the ability to expose geospatial data stored as CF (Climate and Forecast) compliant netCDF files via Open Geospatial Consortium (OGC) standards including Web Map Service (WMS) and Web Coverage Service (WCS). These open standards provide a method of access to remotely hosted datasets allowing users to only download a subset of a much larger dataset, and so go some way to addressing the data download volume issue. Users can make use of locally installed Geographic Information System (GIS) software to access these services to perform their own analysis of the data provided, but this requires a level of knowledge of how to install and use the software. The open source software ncWMS (Blower, 2012) is a web-based software that consumes a WMS endpoint to provide the user with a method of visualising data without having to install software or download an entire dataset. A fairly simple user interface affords the user a method to select an indicator served by the WMS endpoint and to view the data from the selected time step; it is then possible to navigate temporally and spatially using the controls provided.
The use of the both THREDDS and ncWMS require that the software is configured to serve the data by a system administrator before any data can be accessed, and both offer relatively simple catalogue facilities; data are listed as defined by the system administrator with little capacity to add meta data or search capability. This requires that the user knows where to find the data in the first instance, as well as have the necessary skills to make coherent use of the OGC standards.
During the OPEC (Operational Ecology) Project, an EU funded FP7 project, considerable effort was applied to create a web-based tool that provided a single place to discover, visualise and analyse remotely hosted data. This software provided a basic catalogue function allowing users to find data hosted in multiple locations from a single interface. Once the user had identified the relevant dataset it could be explored visually both spatially and temporally using just a standard web browser; there was no software to install, and users only downloaded image tiles in JEPG format from a single time slice for their region of interest. Data was presented using the open source mapping library OpenLayers.
This web-based GIS software provided some simple analysis tools; users could identify a region of interest by drawing a polygon on the map to produce a time series plot that showed the mean, minimum and maximum values of the selected indicator within selected geometry.
WP7 Planned progress beyond the state of the art
The primary aim of Work Package 7 (WP7) was to provide a suite of online services and advanced tools to provide an interface to the data being produced in WP3 (Earth Observations - Combining and improving EO processing techniques) and WP5 (Global water resources reanalysis). These tools should be freely available to use both within the project and for users wishing to use them with alternative data sources. One of the primary things that we wanted to focus on was to offer the ability to share a view of the data with other users, whether the users are sat together or geographically very distant.
The foundation of this work would build upon the open source web-based GIS software produced by the OPEC project to produce the Water Cycle Integrator (WCI) portal. The existing code base would be extended to provide additional features, for example, to offer a greater range of analysis tools and plot types, and the ability to download subsets of a larger dataset. We sought to improve data discoverability by providing better search capabilities, and to offer a greater level of metadata to offer a more complete picture of the data.
The planned work would involve increasing the capacity and resilience of the THREDDS Data server to ensure data availability, speed of access, and the ability to serve a great number of concurrent users.
Alongside these developments we would engage with the user community to gather feedback and to identify features or customisations that could be made to improve the overall solution. The tools provided would be delivered with support from training materials to educate users in how to get the best from them.
WP7 Project advances
Work early in the project focussed on improving access to more resilient data services. The THREDDS Data Server software developed by Unidata would form the foundation of our data access services offering access via WMS and WCS primarily, but also OPeNDAP for more advanced users to download data in their preferred scripting language. In the project a single THREDDS server had been deployed make data available and this presented a single point of failure that we were keen to eliminate. A cluster of four identical physical servers was deployed each with its own instance of THREDDS running, and in front of this we established a load balancer to spread data requests evenly across each of the four servers. This balancing of load ensured continual data provision during periods when one or more of servers would be down for maintenance or software upgrades, but importantly it offered a much greater capacity to handle user requests.
Later in the project (around month 24) we began the process of moving from physical servers to virtual machines hosted within our own cloud environment. This offered even greater flexibility as it allowed us to assign more computing resource to the THREDDS cluster as demand required. The single load balancer was replaced with a high availability load balancer cluster of two virtual machines.
There are a range of tools available to facilitate online meetings, including the likes of WebEx, Skype and GoToMeeting. These tools provide the ability for people in disparate geographic locations to meet online with audio and video inputs, and in many cases screen sharing is possible too. During the meeting, a presenter can share their screen to demonstrate something visual but at the end of the meeting that shared view disappears and the attendees have no further interaction with it. The aim of the collaboration tools developed within this work package was to address this problem by providing a mechanism where two or more people could share a common view of the WCI portal to examine and discuss the data presented. At the end of the meeting all users would disconnect but they could continue their interactions with the visualisation tool independently of the meeting.
To achieve this aim, users are first required to login using Google’s OAuth service; this is purely to identify them by name. The meeting organiser creates a virtual room and is given a reference number, and at this point they can invite as many people as required to join them in the virtual room. Each person who joins the meeting starts an instance of the WCI portal in their own browser. Every interaction that the meeting organiser (or the assigned presenter) has with the WCI portal triggers an event; details of this event are sent via a Web Socket to the host server which is then relayed to all participants in the virtual room. Each participants’ instance repeats the event triggered by the presenter so that everyone in the virtual room is looking at the exact same view of the data. There are also audio and video conferencing features which utilise WebRTC technology; this technology is still in its infancy but allows users to use just a modern web browser to share audio and video in real time and without the need for any other installed software, which is not the case with most other audio/video conferencing tools.
These collaboration tools have been successfully and effectively used to deliver training to users in a class room environment. A teacher can create a virtual room, invite all the students to join them, and then demonstrate the tool and the data it provides access to. The teacher can pass the role of presenter to a student and ask them to perform a task to gauge the student’s understanding. The collaboration tools have also been very helpful when providing support to users; the development team can connect directly with a user and show them how to use a particular element of the WCI portal.
The analysis tools available from development in the preceding project were fairly simple and offered a very static output; it was possible to create a time series plot that was presented as a PNG to download. Developments within the project have extended the range of analysis that it is possible to do online and made the outputs interactive.
Prior to these developments users would have had to download the data that they were interested in, which in itself could be problematic due to the data volumes involved, before using a statistical analysis package (e.g. Matlab) or their preferred scripting language to manipulate the data. However, it is now possible to identify a region of interest, either by drawing a polygon on the map, or uploading a geometry file, or selecting a previously saved geometry; with a region selected, the user identifies the time range that they are concerned with by entering start and end date from the available range, and finally selects from number of interactive plot types. Plot types include time series, Hovmöller plot, scatter, geographic extraction (to provide an average over time within the selected area) or animation. Each of these plots are created on the server from the raw data and just the final result is downloaded by the user. The resulting plots are interactive so the user can move the axis and scroll and zoom within a plot to examine a trend or anomaly in greater detail. Animation plots are created as MP4 videos that can be downloaded for offline viewing or inclusion within a presentation.
The main impact of the eartH2Observe project is the enhancement of the use and exploitation of Earth Observations for water resources management through increasing the availability and accessibility (open access) of these data at the local, continental and global scales. This in turn, directly impacts water resources management and planning, since these data support robust water resources assessments and characterization. The overall project impacts, based on the global project achievements, are presented below. Each list item starts with the E2O achievement followed by the potential impacts.
Achievement: Provision of comprehensive multi-scale (local, regional, continental and global) water resources observations and datasets.
- Support globally consistent assessment of the water resources, and thus contribute to efficient water management and decision making at multiple levels (from local to global).
- Increased availability of data necessary for water resources analysis (especially the accessibility for data-poor regions, and transboundary basins
Achievement: Integration of available global earth observations (EO), in-situ datasets and models.
- These data allow for inter-comparison, better validation and thus better estimates
Achievement: Construction of a global Water Resources Re-analysis dataset (WRR) of significant length (30 years), based on common forcing datasets and an ensemble of models.
- These data allow for improved insights on the full extent of available water and existing pressures on global water resources, in all parts of the water cycle.
- They also provide a baseline and a benchmark for assessing past and future trends. By post-processing these data, a variety of indicators can be derived (e.g. on drought, floods, etc.)
- Additionally, the WRR can support international agreements on water management and European water policies (e.g. the EU WFD, the Flood Directive, Communication on Water Scarcity & Drought, the EU Water Accounts, the EDO, etc.)
Achievement: Testing of new EO data sources, extension of existing processing algorithms and combination data from multiple satellite missions.
- Contribution to the improvement of the overall resolution and reliability of EO data, as well as their usability and uptake.
- The developed retrieval and optimization algorithms are generic in nature to ensure their compatibility with satellite missions launched in the near future.
Achievement: Verification and demonstration of the usability and operational value of the developed data in 7 international case-studies (Spain, Morocco, Estonia, Bangladesh, Ethiopia, Colombia, Australia/New Zealand).
- At the global scale, the data produced by the project have been validated in international case studies, and thus informed conclusions about their usability, added-value, etc., are available for other users. Furthermore, it is ensured that the data can be made operational.
- At the local scale, stakeholders and end-users in the case study have gained valuable knowledge on how to improve the efficiency of regional water distribution, and training on using on state-of-the-art tools in their planning and decision making processes.
- The case studies further serve as exemplars for the application of the datasets and models in other areas and/or (developing) countries.
Achievement: Assessment of error propagation through large scale water resource modeling, using in-situ data from data-rich validation sites.
- Improvements of the EO datasets, availability of datasets to end-users together with their associated error characteristics so that they have metrics of uncertainty.
Achievement: Public dissemination (open access) all of the produced datasets through the Water Cycle Integrator (WCI) portal.
- Increased availability of water resources information, easily accessible, freely retrievable.
- Allow (through the WCI) for advanced data visualization, analysis and exchange
Achievement: Contribution to the existing GEOSS water cycle platforms and communities
- The E2O results will become part of the GEOSS Data-CORE encyclopedia, connecting to the GEOSS Water Cycle Integrator (WCI) initiatives and openly providing and sharing the datasets developed in the project. Furthermore the data supports the SDG’s via GEO.
Achievement: Active dissemination of all project results and data to multiples target audiences.
- Implementation of training, exploitation and networking activities.
- Contribute to capacity building in the wider field of EO.
- Exploitation and promotion of EO datasets, including new applications and datasets.
- Contribute to the science-to-policy and science-to-business interfacing by networking with stakeholders and SMEs.
- Help SMEs to develop and improve services and products in the domain of environmental assessment and monitoring using EO technology.
An assessment (based on questionnaires filled by the participants) of the added-value and potential impact of the E2O overall achievements and their contribution to the field of “earth observation and water resources” has been carried out during the 11th European GEO Workshop (Helsinki, 19-21 June 2017), within the E2O break-out session (on 20/06/2017) “Using Global water resources data to derive actionable information at the local scale”. The following E2O achievements ranked-up by the workshop participants (E2O achievements Added-value/ Potential Impact, max score =5)
- Open Access to Data – E2O Water Cycle Integrator (WCI): 4,89
- Data Harmonization: 4,13
- Emerging Business & Downstream Applications linked to GEOSS: 4,13
- Downscaling and integration of EO Data in local hydrological models: 4,11
- Integration of data from multiple satellite missions: 4,00
- Policy analysis, hotspot identification, common Indicators: 4,00
- Higher resolutions: 3,89
- Long -time series, baseline Reference Dataset (>30 yrs): 3,89
- Improved coverage: 3,63
- Uncertainty analysis: 3,13
- Ensemble-models’ results in Water Resources Reanalysis (WRR): 2,88
- Common forcing dataset, forcing models’ improvements: 2,75
- Modeling error propagation analysis: 2,75
E2O Exploitable products
The specific project impacts and the exploitation results, based on the specific E2O exploitable foreground (specific products and outputs), are presented. These exploitable products are namely the following:
1. E2O Water Cycle Integrator (WCI) Portal and related Datasets; 2. Snow Portal; 3. Water Resources Reanalysis (WRR2 and ENSEMBLE WRR2 datasets); 4. E2O Down-scaling Tools for WRR2; 5. Global Hydrological, Land Surface and Water Balance Models; 6. WaterWorld Policy Support System; 7. Calibrated river basins models and datasets; 8. Precipitation estimates datasets (CDRD, PNRPM 183-WSL, MSWEP); 9. SUR –Synthetic Aperture Radar Precipitation Estimation algorithm (SIDOC-MREA) ; 10. SAFRAN meteorological analysis for Spain; 11. E2O e-learning platform
1. E2O Water Cycle Integrator (WCI) Portal and related Datasets
Product brief description: The Water Cycle Integrator (WCI) Portal (https://wci.earth2observe.eu/) is an open source web-based visualization and analysis tool that offers the ability to view, analyze and download any of the data produced by the project partners (i.e. hydrometeorlogical and water resources related data).
Type: 1-General advancement of knowledge; 2-Commercial exploitation of R&D results (for the software); 3-Exploitation of R&D results via standards
End-users: Researchers, scientists, academia, SMEs, consultants, policy and decision-makers and the general public which seek information (data and related statistics) of the state of global water resources or specific data on meteorological, hydrological and water resources variables.
Potential Impact: The expected impact of the WCI is multifold: open access to state-of-the-art global water resources datasets necessary for multiple level users from different disciplines, support advanced data analysis and visualization, support online collaboration, serve research and educational purposes, support policy analysis and decision-making, allow for the exploitation of earth observation data by the consulting and business communities.
The tool can be further used in a variety of scenarios where geospatial data are involved and there’s a need to view it quickly and simply, for example, training courses on data usage and data access, in schools for education on GIS.
Exploitation means & pathways, and further developments: The E2O WCI portal will be up and running for 5 more years after the project end, i.e. until 31/12/2022. The use of the portal will be pursued by the project partners, but also by other interested groups from the research, policy and business communities, through the partners, the stakeholders in the case study areas, and the E2O website and after-life dissemination activities. The WCI has already been demonstrated to the EEA (Water Accounts purposes), the EC DG Environment and SMEs during the event “E2O Marketplace of Ideas” (in Delft, on 01/12/2017) who demonstrated an increased interest in using it, and also attempted to identify (the SMEs) further applications which could be developed using data from the E2O WCI. Actions to identify possible funding for further maintaining and updating the E2O WCI Portal, and discussions with GEOSS have been established, also looking into the potential for turning the E2O WCI Portal into a Copernicus related service. It has to be noticed that the soil moisture datasets developed/ improved within E2O (by partner TU WIEN) have led to the exploitation of this specific product (soil moisture) and a related agreement for a Copernicus service (by TU Wien) has been established. The WCI software itself will be used and further developed in projects funded by e.g. the Copernicus programme and offered on a commercial basis, so its ongoing development is assured to some degree.
2. Snow Portal
Product brief description: The aim of the Snow Portal (http://snow.gisat.cz/intro/#intro) is to support analyses and assessment of the usability of different types of optical satellite imagery for snow monitoring purposes
Type: 1-General advancement of knowledge; 3-Exploitation of R&D results via standards; 4-Exploitation of results through EU policies
End-users: European bodies, researchers, scientists, academia, SMEs, consultants, policy and decision-makers and the general public which seek information about the data availability of cloud-free vs cloud-covered areas over Europe from selected optical satellite imagery.
Potential Impact: The portal shall support the European Environment Agency (EEA) in its decisions related to the future European snow & ice cover monitoring service within the Copernicus programme. GISAT has ongoing collaboration with the EEA. The use of the portal will be also pursued by the project partners, but also by other interested groups. The Snow Portal has already been demonstrated to the EC DG Environment and SMEs during the event “E2O Marketplace of Ideas” (in Delft, on 01/12/2017) who revealed an interest in using it.
Exploitation means & pathways, and further developments: The expected impact of Snow Portal is to improve operational snow monitoring at the European scale. It is mainly foreseen in supporting the European Environment Agency (EEA) in their decisions related to designing a new snow & ice monitoring service, and to serve scientists and others in their research. Further development is potentially foreseen mainly based on the feedback from the EEA and their further needs.
3. Water Resources Reanalysis (WRR2 and ENSEMBLE WRR2 datasets)
Product brief description: The E2O project developed a global Water Resources Re-analysis (WRR) based on state-of-the-art meteorological re-analysis, improved with earth observations and extended with output from hydrological and land surface models
Type: 1-General advancement of knowledge; 3-Exploitation of R&D results via standards
End-users: Researchers, scientists, academia, SMEs, consultants, policy and decision-makers and the general public which seek information (data and related statistics) of the state of global water resources or specific data on meteorological, hydrological and water resources variables. They can be used for multiple purposes, including: Meteorological and hydrological monitoring and forecasting applications; Regional and global scale modelling; Enhanced representation of terrestrial water resources (e.g. soil and snow reservoirs); Water resources assessment with variability due to the forcing; State-of-the-art benchmark for future model developments;; Water management and water science; Water policy advice
Potential Impact: Multi-model reanalysis of the state of the surface water storage and fluxes provide an ensemble that is not dependent on a single model. The multi-model approach has been a recurrent approach in several international projects (e.g. GSWP2, Dirmeyer et al. 2006; EU-WATCH, Harding et al. 2011), and in E2O we succeeded in generating a global version of the Water Resource Reanalysis based on a set of different land surface and hydrological models simulations with their current modeling system with a controlled modeling protocol. Thus, the expected impact of the WRR2 products is multifold: open access to state-of-the-art global water resources reanalysis datasets necessary for multiple level users from different disciplines, support advanced data analysis and comparison, provide a benchmark for uncertainty analysis in water resources modeling and allow to evaluate the quality of model simulations, serve research and educational purposes, support policy analysis and decision-making, allow for the exploitation of processed earth observation hydrological data by the consulting and business communities. The WRR2 provides a robust, bias-proofed product that can be used in downstream applications of water resources, including policy-relevant decision making. Additionally, it constitutes a benchmark enabling to evaluate improvements (including the value of earth observations to improve the representation of hydrological processes) even after the E2O project lifetime.
Exploitation means & pathways, and further developments: The WRR2 datasets are hosted by the E2O WCI Portal and will thus be available via the WCI for at least 5 more years. The use of the WRR2 products will be pursued by the project partners, but also by other interested groups from the research, policy and business communities. The WRR2 products already been demonstrated to various stakeholders at the local level, at the global an EU level, and to the EC DG Environment and SMEs during who expressed an increased interest in using the, and also attempted to identify (the SMEs) further applications which could be developed using these datasets. Actions to identify possible funding for further updating and extending the WRR2, and discussions with GEOSS have been established, also looking into the potential for developing a Tier-3 WRR. The E2O partners who own the models used to produce the WRR2 will continue developing and improving these models through other projects and funding (e.g. increasing resolution, merged products based on a Tier-3 WRR by a larger number of ensemble members). New model developments and improvements can lead to a new Water Resources Reanalysis (WRR tier-3) throughout the community created within E2O.
4. E2O Down-scaling Tools for WRR2
Product brief description: The E2O Down-Scaling Tools consists of a number of python programs and procedures that facilitate local application of the E2O global Water Resources Reanalysis. The tools can connect directly to the E2O WCI Portal and save (resampled) data to a local computer for further analysis or direct application. The tools downscale the WRR1 and WRR2 datasets based on a higher resolution digital elevation model of the area of interest.
Type: 1-General advancement of knowledge
End-users: Researchers, scientists (hydrologists), SMEs, consultants, water resources managers and planner active in the field of hydro-meteorological analysis at the river basin scale, that require high-resolution data of the state of global water resources or specific data.
Potential Impact: Publicly available, easily accessible and user friendly down-scaling tools necessary for multiple level users from different disciplines who require high-resolution data of the state of global water resources or specific data on meteorological, hydrological and water resources variables. In combination with the E2O WRR2 (and WRR1) dataset, these tools facilitate and further boost the users’ capacity to perform hydro-meteorological analysis at the river basin scale.
Exploitation means & pathways, and further developments: The use of the E2O Down-scaling Tools will be pursued by the project partners, but also by other interested groups. An e-course is associated with the E2O Down-scaling Tools, available on the e-learning platform on the E2O website. The Tools have already been demonstrated to SMEs who expressed an interest in using them in combination with the E2O WRR2 datasets. The use of the E2O Down-scaling Tools will be pursued by the E2O partner DELTARES (i.e. the developer), but they can also serve as educational material to other E2O academic partners (to be used in their classrooms).
5. Hydrological, Land Surface and Water Balance Models
Product brief description: In order to run a global Water Resources Reanalysis and produce the various WRR datasets ten hydrological, land use and water balance models have been used:
HTESSEL (by ECMWF); PCR-GLOBWB (by UU); SURFEX-TRIP (by Meteo-France); JULES (by CEH); LISFLOOD (by JRC); ORCHIDEE (by CNRS); WaterGAP3 (by UniKassel); W3 (by ANU/CSIRO); WaterWorld (by AmbioTEK)
Type: 1-General advancement of knowledge; 2-Commercial exploitation of R&D results
End-users: Researchers, scientists, academia, SMEs, consultants, water managers, river basin authorities, river basin commissions, operational flood forecasting authorities, policy and decision-makers, who wish to model the hydrological cycle, the water balance, the land surface interactions, etc. The models have a range of potential applications, such as: enhanced representation of the terrestrial water resources (e.g. soil and snow reservoirs), improved diurnal to seasonal-range forecasting, monitoring and forecasting of water resources (flood, droughts - already operational under Copernicus).
Potential Impact: The expected impact of the models used and improved within E2O is mostly brought on the scientific and research community (i.e. availability of better, more robust modeling products, availability of a benchmark for inter-comparison and uncertainty analysis, identification of RDI areas which call for further development), but also on local and global stakeholders (water manager, river basin organizations, water policy and decision makers) who have a wide option of LSM and Hydrological Modeling tools, along with a validation/ verification analysis to use for their own purposes (water management and/or water policy analysis).
Exploitation means & pathways, and further developments: The use of the models will be pursued by the project partners, but also by other interested groups. Further developments of each model individually have been identified and will be pursued by the corresponding model developer. Actions to identify possible funding for further developing and updating the various models will be pursued by each model’s beneficiary partner. At a higher project level, we are looking into the potential for turning the E2O WCI Portal into a Copernicus related service. In this direction, the outputs of the models (WRR products/ datasets for an extended period beyond 2014) will need to be update, which, in turn, means that the models will also need to be re-used and upgraded.
6. WaterWorld Policy Support System
Product brief description: WaterWorld is a web-based global water resources policy support system that has been developed at 1km and 1ha resolution over the past 20+ years and is aimed at non-technical policy decision makers or users on low bandwidth connections. The new version incorporates detailed spatial datasets at 10-square km resolution for the entire world, spatial models for biophysical and socio-economic processes along with scenarios for climate, land use and economic change.
Type: 1-General advancement of knowledge; 2-Commercial exploitation of R&D results
End-users: Researchers, scientists, academia, SMEs, consultants, policy and decision-makers and the general public which seek information (data and related statistics) of the state of global water resources, and the impact of scenarios and interventions such as climate change and land use change on the water resources.
Potential Impact: multifold: open access to state-of-the-art global water resources modeling necessary for multiple level decision makers, advanced data analysis and visualization, research and educational purposes, support policy analysis and decision-making, allow for the exploitation of EO data.
Exploitation means & pathways, and further developments: The development of WaterWorld at 1km and 1ha resolution has taken place over 20+ years in close collaboration with Dept Geography, Kings College London. In E2O the same capabilities have been scaled up to 10km resolution which is nearer the scale demanded by users of remotely sensed data. This version is being used and further developed in the H2020 Helix projects. The 1ha and 1km resolution version is offered on a commercial basis to commercial users, and free of charge to non-commercial users, so its ongoing development is assured to some degree. New features and functionality will be added as required by those projects that make use of the software.
7. Calibrated river basin models and datasets
Product brief description: A suite of local and global hydrological models that has been calibrated for large river basins in six countries. The models have been calibrated and validated using in-situ available data and provide the best possible estimates of the river basins hydrology.
Type: 1-General advancement of knowledge
End-users: Hydrological modelers, water manager, river basin authorities, local stakeholders involved in water planning and decision-making, policy makers, universities and research institutes, SMEs and consultants.
Potential Impact: The local applications developed within E2O, with direct end-user interaction in several international case studies (including non-European developing countries) support the provision water management advice to the local public authorities and citizens that seek an improved distribution of freshwater at the regional and local level. In each case study the developed datasets and models can be used to assess local water resources availability and the severity of the water related problems (e.g. floods, water shortage, water pollution). In turn, possible interventions and adaptation measures in the river basins can be designed and implemented based on best scientific knowledge and local policy input. Extrapolation and/ or replication of the work in other areas is feasible, since the E2O case studies have verified/ validate the local applicability of the EO products and can serve as exemplars for the application of the datasets and models in other (developing) countries.
Exploitation means & pathways, and further developments: The global models and related datasets are available through the E2O Water Cycle Integrator (WCI) Portal. The locally calibrated river basin models are available through the partner institute of the respective case study. The local partners will pursue model’s improvements, as well as maintaining and strengthening their links with the local decision-makers and water managers for a continuous uptake of the results and for better linking science to policy.
8. Precipitation estimates datasets (CDRD, PNRPM 183-WSL, MSWEP)
Product brief description: Various precipitation estimates datasets as follows:
▪ Cloud Dynamics Radiation Database (CDRD) v1 and v2 estimates: provides precipitation estimates from conical scanning microwave radiometer SSMIS covering the Europe-Africa region
▪ Passive Microwave Neural-network Precipitation Retrieval (PNPR, PNPR2) estimates: provide precipitation estimates from across-track scanning microwave radiometers (AMSU, MHS, and ATMS) covering the Europe-Africa region.
▪ Water vapour Strong Lines at 183 GHz (183-WSL) precipitation intensity estimates: provides precipitation estimates from across-track scanning microwave radiometers (AMSU, MHS) with a global coverage
▪ MSWEP global precipitation product at 0.25° resolution: provides an unprecedented global precipitation product encompassing all possible sources from the ground and from satellite remote sensing
Type: 1-General advancement of knowledge; 2-Exploitation of R&D results for operational hydrology and water management
End-users: Possible users for the algorithms and corresponding datasets are the meteorological services, hydrological modelers, water manager, universities and research institutes, SMEs and consultants who develop relevant services, insurance companies, health services.
Potential Impact: The various precipitation estimates dataset have applications in multiple fields, such as precipitation monitoring (including remote areas and oceans), extreme events monitoring, hydrological modeling. As these products are improved estimates and of adequate resolutions they can be used for better hydrometeorological analysis and assessments. More particularly, the MSWEP product features a truly global coverage (including ocean areas) at 3-hourly 0.1° resolution (other satellite-based datasets, such as TMPA 3B42, are limited to latitudes <50/60°), offering a consistent precipitation record from 1979 until the near present, enabling trend and drought assessments.
Exploitation means & pathways, and further developments: The various precipitation estimates datasets take advantage of the complementary strengths of gauge-, satellite-, and reanalysis-based data to provide reliable precipitation estimates over the globe. To this end, they can be further exploited by the insurance field in applications for crop damage and flood/drought risk assessment. Additional applications are related to the evaluation of the transmission/diffusion of diseases (e.g. malaria), whose dynamics is substantially determined by the spatial and seasonal patterns of rainfall. The E2O partner CNR who developed these products will pursue further developments.
9. SUR – Synthetic Aperature Radar Precipitation Estimation algorithm (SIDOC-MREA)
Product brief description: The SIDOC-MREA algorithm allows to detected areas affected by intense precipitations or flood in X-Band Synthetic Aperture Radar images. Moreover, it allows estimating the precipitation rate for the areas affected by precipitations. The added-value of X-SAR precipitation maps is the very high spatial resolution (of the order of hundreds of meters) and the possibility of observing impervious areas.
Type: 1-General advancement of knowledge; 2-Commercial exploitation of R&D results
End-users: The SIDOC-MREA framework can be used by researchers, scientists, academia interested in precipitation measurements, weather monitoring and modeling and flood monitoring or interested in SAR ground products unaffected by precipitations. The use by SMEs and other non-technical users requires the development of a closed-form user-friendly tool, foresight by SUR.
Potential Impact: SIDOC-MREA impact is expected to be relevant especially in the field of climate modeling where the high spatial resolution can be particularly appreciate. Also, the precipitation-masking feature is expected to have a relevant impact.
Exploitation means & pathways, and further developments: The SIDOC-MREA framework allows exploiting the sensitivity of X-Band SARs to intense precipitations in order to derive precipitations maps at a very high spatial resolution. Moreover, satellite X-SARs allows the observation of impervious areas, independently by solar irradiance conditions. Finally, the detection and classification features allow the detection and monitoring of flood areas, and to ground product users the masking of areas affected by precipitations. SUR foresights the possibility to further improve the framework and realize a commercial tool, as well as the patent registration of the developed tool, once terminated and validated. SIDOC-MREA nowadays applies only to land; SUR has planned to extend to sea surfaces. SUR has also planned to validate the framework in other study cases, both real and simulated, using the developed Forward Model, and to realize a commercial user-friendly tool.
10. SAFRAN meteorological analysis for Spain
Product brief description: SAFRAN meteorological analysis for Spain is a gridded dataset of all the meteorological variables necessary to force a physically based land-surface model. The dataset has a spatial resolution of 5 km, a temporal resolution of 1h, it spans from 1979 to 2014 and covers mainland Spain and the Balearic Islands.
Type: 1-General advancement of knowledge
End-users: Meteorologists, climatologists, hydrologists agricultural engineers, biologists, water managers, decision-makers, policy-makers, SMEs and consultants who develop services/ conduct relevant studies, national and regional authorities
Potential Impact: The purpose of the Spanish SAFRAN dataset is providing data that allows running any kind of physical land-surface model. Such a dataset was inexistent in Spain and its existence allows many future scientific developments.
Exploitation means & pathways, and further developments: The SAFRAN dataset can be exploited by any research institution that owns a model that needs such data as forcing. The dataset is restricted to research purposes and it is subject to the HyMeX Data Polilcy. The dataset is mature, however future improvements are possible in order to improve the quality of the hourly data for some variables. These data opens new research possibilities in Spain that were not possible until now.
11. E2O e-learning platform
Product brief description: The E2O e-learning platform (http://www.earth2observe.eu/?page_id=5679) is a public area of the E2O website, accessed by all interested users, and provides online training and educational material, structured under 5 specific e-courses:
Hydrological modeling down-scaling tool (DELTARES ©); Training on using the E2O WCI Data Portal (PML ©); Using WaterWorld to Understand Current and Future Water Resources Indices (AmbioTEK ©); Long-term hydrological monitoring with active and passive microwave missions (TU WIEN ©); Using the E2O data portal to analyze drought indicators (by IHE Delft ©)
Type: 1-General advancement of knowledge
End-users: The E2O e-learning platform has been set up to support training and educational activities towards all the target groups of the project. It can be used by college students and professors, researchers and scientists, SMEs, consultants, practitioners and water managers, regional and global stakeholders, policy analysts, decision makers, NGOs (depending on the selected e-course).
Potential Impact: The expected impact of the E2O e-learning platform is multifold: educate various users’ groups on the application of the developed E2O tools, build capacity, provide training facilities, foster the uptake of the E2O products and the creation of communities of practice, contribute to the dissemination of the scientific knowledge gained through the project research activities.
Exploitation means & pathways, and further developments: The E2O e-learning platform will be up and running for 5 more years after the project end (hosted by the E2O website), i.e. until 31/12/2022. The use of the e-learning platform will be pursued by the project partners, but also by other interested groups. Many project partners are from the academia, teaching courses and interacting with students, and it is thus expected that they will also advertise the E2O courses in their classrooms/ students’ communities where relevant. A message will also be displayed in the WCI portal prompting the users to the available e-courses of relevance to WCI training, data assimilation etc. AmbioTEK will also promote the available training facility to its wide network of policy stakeholders and DELTARES will use (and possibly) extend the training materials in interaction with their clients. Should the E2O partners wish to develop and upload additional e-courses SEVEN will undertake this task as part of the future exploitation of the e-platform, with the vision to expand and enrich it.
Additional exploitation achievements pursued by project partners
TU-Wien has worked to the C3S_312a Lot 7 Copernicus Climate Change Services: Essential Climate Variable products derived from observations
Lot 7 – Soil Moisture. The objective is to contribute to ECMWF Climate Data Store (CDS) by providing open, free access, quality-assured, state-of-the-art soil moisture in “near-real-time”. The service is operational since July 2017.
Estellus has delivered the GIEMS-D3 dataset to several research groups, for both regional and global studies. The new GIEMS database will be proposed to the climate community. In particular, the latest years to present time are waited for by the climate community to validate some hypothesis about the emissions by wetlands.
An international action is being organized, gathering the major inundation database producers (D. Yamazaki, B. Lehner, J.F. Pekel and the Estellus group), to build a reference dataset to be delivered to the scientific community. The two most important components will be the description of the open waters from VIS (Landsat data) and the flooding from passive microwave (GIEMS-D3) observations.
The method developed by Deltares to quickly and interactively extract surface water from Landsat data was picked up by the NASA SERVIR-MEKONG team and further developed into an on-line app that is used by stakeholders in the region (https://servir.adpc.net/tools/surface-water-mapping-tool).
Several partners have cooperated and submitted calls for proposals to different funding mechanisms in pursue of extending the E2O work
Main Dissemination Activities
The main dissemination activities realized within the E2O are summarized below.
Dissemination and/or communication activity: E2O website
Target Group: Scientific, policy and business communities, NGOs and the general public
Description: Dynamic website
Impact indicators: No. of Sessions: 13,40; % New Sessions: 64.91%; Users: 8,699; Pageviews: 34,336; Bounce rate: 55.90%
Dissemination and/or communication activity: E2O Water Cycle Integrator (WCI)
Target Group: Scientific, policy and business communities, NGOs and the general public
Description: Data portal
Impact indicators: Total data volume stored 17.7TB from 19 different providers offering 1,158 unique data variables covering the period from January 1979 through to December 2014.
Average 255 users per month.; 255 pageview per month
Dissemination and/or communication activity: E2O Facebook page
Target Group: Scientific, policy and business communities, NGOs and the general public
Description: E2O dedicated page on Facebook (open to all public)
Impact indicators: Total likes: 2,449; Total number of followers: 2,448; Most popular post: 3,182 people attracted/1,184 likes & comments
Dissemination and/or communication activity: E2O ResearchGate page
Target Group: Academia, scientists, researchers, practitioners
Description: E2O dedicated page on ResearchGate
Impact indicators: Total number of followers: 92; Reads: 879; References: 743
Dissemination and/or communication activity: Promotional Material
Target Group: Scientific, policy and business communities, NGOs and the general public
Description: Various promotional and dissemination material
Impact indicators: Leaflets: 1000; Posters: 51, Newsletters: 5
Dissemination and/or communication activity: Policy Briefs
Target Group: Local and national policy makers, decision-makers, water managers, EU and International policy makers
Description: Policy-related and targeted briefs
Impact indicators: Policy briefs: 1; Science-policy letters: 1; Policy-relevant Fact Sheets on the outputs of the E2O case studies: 7
Dissemination and/or communication activity: Workshops and side-events organized by the E2O
Target Group: Scientists, local and/or national stakeholders, water managers, river basin authorities, water planners and decision-makers, EU policy makers, SMEs, water utilities
Description: Case Study Workshops (19); E2O Sessions/ side-events in international or EU Conferences/ Workshops/ Symposia (16)
Impact indicators: 419 participants in 19 Case Study Workshops; 513 participants in 16 E2O Sessions/ side-events
Dissemination and/or communication activity: Training and Educations
Target Group: Scientific community (higher education, research, PhD students,), practitioners, water managers, SMEs, NGOs, regional and global stakeholders, policy analysts, decision makers
Description: Trainings/ Webinars (6); E2O e-learning platform
The WCI portal is used in courses at TU Twente, Hogeschool Larestein and IHE Delft
Impact indicators: 328 participants in 6 trainings/ webinars; 5 courses available on the E2O e-learning platform
Dissemination and/or communication activity: E2O Final Conference and E2O MarketPlace of Ideas (MoI)
Target Group: Scientists, SMEs, policy makers (E2O partners, AB members, SMEs, DG Environment)
Description: Final Conference of the E2O project, combined back-to back with a Brokerage exploitation-oriented event (MoI)
Impact indicators: 50 participants
Dissemination and/or communication activity: Scientific Publications
Target Group: Academia, scientists & researchers, policy community, water managers, practitioners
Description: Peer-reviewed publications in Journals; Presentations & publications in Conferences; Scientific articles; Special Issues; Doctoral Thesis
Impact indicators: Peer-reviewed publications in Journals: 110 (of which 85 are open access); Presentations & Publications in Conferences: 136 in 74 different conferences/ workshops; Scientific articles: 4; Special Issues: 3; Doctoral Thesis: 3 finalized, 20 more PhD students in progress
List of Websites:
contact details project co-ordinator: firstname.lastname@example.org
Grant agreement ID: 603608
1 January 2014
31 December 2017
€ 11 344 199,01
€ 8 869 787
Deliverables not available
Grant agreement ID: 603608
1 January 2014
31 December 2017
€ 11 344 199,01
€ 8 869 787
Grant agreement ID: 603608
1 January 2014
31 December 2017
€ 11 344 199,01
€ 8 869 787