INTOGENER (INTegration of EO data and GNSS-R signals for ENERgy applications) Demo is a project aiming at demonstrating the operational capabilities of a water flow monitoring and prediction system aimed at hydropower production and water management organizations.
The INTOGENER system combines different technologies: satellite-based Earth Observation (EO) data, in-situ information based on reflected Global Navigation Satellite Signals measured from remote places and transferred by satellite data links, and finally, assimilation of these data into a distributed hydrological model water flow forecast.
INTOGENER demo campaign will take place in Chile, where the combination of technologies (EO and non EO data) will ensure periodical information based on physical observations rather than on statistics, as it is currently the case.
The INTOGENER Demonstration Project is the result of a feasibility study which was successfully concluded in 2011.
Objectives of the service
Having an accurate forecast of available water flow is crucial for hydropower companies and water management authorities in order to optimize operations and improve efficiency.
In the frame of the ESA Integrated Applications Promotion (IAP) programme (ARTES element 20), Starlab has just started the second phase of the INTOGENER project, aiming at the development and demonstration of a new, operational, water-flow forecast concept, based on the assimilation of near real time measurements of geophysical parameters into a hydrological model.
The system makes use of measurements derived both from Earth Observation satellites (SAR and Optical instruments for the retrieval of snow cover and temperature maps), and from in-situ sensors (for temperature, precipitations, soil moisture, solar radiation, water level). Satellite links are used to transfer information from the field (remote mountainous areas) to the processing centre.
During a first phase of the activity, one basin in the Chilean Andes has already been equipped, and sound predictions have been obtained. In the next two years, in partnership with Endesa and the Pontificia Universidad CatÃ³lica de Chile (PUC) and with the support of Hispasat and Future Water, Starlab will demonstrate the real-time operational capabilities of the system, as the last step before potential commercialization of the service.
Users and their needs
INTOGENER Demo is targeting water management applications. The more accurate streamflow predictions are, the better water management and energy efficiency can be attained, with the derived economical benefit.
The user area is renewable energy production with a focus on hydroelectric production companies. Starlab has established along the years a strong partnership with Endesa Chile, which is a key player in the region, being the main producers of hydroelectricity in Chile.
A first analysis of User Need has been performed, leading to a set of statements concerning the current status of flow prediction:
- The indicators and methodology used to assess the hydrological uncertainty, i.e. flow prediction, through models are key to obtain valuable results to monitor the plants energy production.
- Current operations models used in water management require hydrological information of water resource availability, particularly during drought events. To monitor these events is also very important for risk management.
- The current hydrological model inputs are based on in situ measurements and statistical analysis based on past events.
- In situ data might sometimes be of difficult retrieval due to the extremely difficult access to the area to be monitored.
- There is a clear need to include EO near real-time data in the prediction methodology and to integrate it into the model.
- Flow prediction tasks are often subcontracted. Statistical Tools are the most common prediction tool in certain areas.
- Flow prediction models used by majors companies can have between 40% and 80% of accuracy.
Service/ system concept
The INTOGENER service will provide flow predictions at certain points of a Chilean hydrological basin. These predictions will be produced by a distributed model that assimilates:
- Water level measurements retrieved from reflected GNSS signals in the main basin's lake, from which the available resource is assessed.
- Temperature, humidity, rainfalls and solar radiation information collected from automated weather stations.
- Snow Cover Area maps from Earth Observation data
A data collection platform (Satellite modem and software) will perform data acquisition and control the GNSS-R instrument by a satellite link.
Space Added Value
The space assets to be used include:
Communication satellites for transmission of in situ data from remote sites (not possible by terrestrial ways in remote areas)
GNSS Satellites for estimation of water level in remote lakes using GNSS-R instrument (Low maintenance, works in any weather, automated, allows large water lever excursion).
Earth Observation satellites to derive:
Snow cover from SAR (RADARSAT) and optical sensors (MODIS). Huge geographical coverage compared to in situ snow routes
Space-based Digital Elevation models. Very large coverage and availability compared to plane based or other methods.
Land Use Maps. Large coverage and availability compared to other methods
The INTOGENER flow prediction concept is a turnkey solution. The quality of the results is independent from the existence of similar climatological conditions in the historical records. It includes state of the art distributed hydrological modelling fed with near real time space and in situ measurement to reflect the true state of the basin. The system can theoretically be adapted to any new geographical area, no matter the remoteness of the sites to be monitored.
In comparison, the existing solutions for water flow predictions only partially cover the user needs. Although a growing number of hydrological models are appearing, their setup, calibration and use require know-how that hydropower companies are generally focusing on more operational and statistical models. Equally, EO solutions for snow parameters are improving and will continue to do so in the future with new satellites and algorithm development breakthroughs. Nevertheless, EO services rarely include flow forecasting at specific points, and those that do are limited to geographical areas and are conducted under very specific funded projects, being service sustainability still a challenge.
Intogener is currently the only pre operational integrated solution using both space and non space assets in a hydrological model to predict water flow.
INTOGENER aims at improving the water flow predictions in remote mountainous catchments were most of the hydroelectric production is located. For this purpose the service concept developed by Starlab uses near real-time geophysical information from in situ and remote sensing space based instruments to feed a hydrological model producing water flow forecast.
The in situ measurements identified as relevant for flow prediction are: temperature, precipitation, solar radiation, water level, soil moisture. Reservoir water level is estimated using a novel technology based on the measurements of Global Navigation Satellite System (such a s GPS and Galileo) signals reflected by the lake surface (GNSS-R). If required, in situ information can be sent by satellite communications, in order to provide real-time data to the model.
The space based Earth Observations identified as relevant for flow prediction are temperature maps and snow cover. Snow cover and temperature maps can be measured by optical instruments, with the limitation of cloud cover that occurs regularly during the melting season, and coarse resolution not suitable for hydrological modelling. Whereas temperature can be completed by in situ measurement, extrapolated and downscaled using a digital elevation model, snow cover is more difficult to downscale and extrapolate. For this reason the proposed service includes the acquisition of SAR images which are cloud immune and have a resolution suitable for the hydrological model. Using state of the art algorithms, optical and SAR images can be fused in order to produce high resolution realistic maps of the snow cover.
The hydrological model is initialized with digital elevation model, land use and soil type maps, and calibrated using historical meteorological and flow data. It is then capable of producing short term (days) to long term (seasons) water flow forecasts at points of interest by taking as starting points an updated state of the basin (from updated geophysical information) and a climatology modulated by El NiÃ±o Southern Oscillation (ENSO) index as prediction of the meteorological variables.
All the information (from in situ instruments, satellite observations, historical data, and model output) is centralized on a secure server running a database as well as services required for data flow and automation. Routines on this server will produce water flow reports, including quality indicators, as well as system health indicators available by the service team.
The output of the service is a complete forecast of the water-flow based on near real time physical observations delivered automatically to the user on a weekly basis. This forecast is computed at points of interest of the basin, that match input variables in the operational model, and therefore usable directly in their management practices.
In areas like Chile, hydro power companies have been using forecasting solutions such as statistical regressions of flow observations, or mathematical simulations with a minimum content of real-time physical information. Recently, a succession of non-average climatic years made an impact on their energy efficiency planning, as the models were not able to predict the water availability.
Moreover, in remote locations with large inflow reservoirs of difficult access and challenging forecasting points, current solutions are limited by the scarce amount of in-situ data, complex operational systems, extreme topographic conditions, and lack of communications systems.
The INTOGENER Demonstration project was kicked-off on May 11, 2012.
On site deployment of the satellite communication system and weather stations was performed successfully during the first week of May 2013.
Some pictures of the installation can be seen below:
The operational service delivered water forecast predications to the user ENDESA on a weekly basis for the melting season 2013 – 2014.
A web site was set-up to allow the user:
to view all the in-situ measurements stored in the database,
to download weekly water flow forecast reports
to visualise weekly water flow forecast reports on-line in the form of charts
to access a series of key performance indicators
to check the health of the system
A screenshot of the web site is shown below.
The service was run operationally for the melting season 2013 – 2014.
- 155 days, from the 25th of September 2013 up to the 26th of February 2014.
- 22 weekly forecasts (Every Wednesday 7am Chilean time) were provided to ENDESA.
- 10 SAR images for SCA computation were used.
- Daily processing of optical images for SCA computation was performed.
- Daily gathering of data from 2 dedicated weather stations (1 per basin), sending data by satellite communications, was performed.
- Daily gathering of data from more than 15 flows and weather stations from the DGA was performed.
The final review took place on May 27, 2014. The project is now completed.
As conclusions of the INTOGENER demonstration project, it can be stated that:
- INTOGENER service has shown good compliance with the operational requirements from the point of view of utilities as well as for the service capacities from a provider point of view.
- Out of 6 forecast points, 3 gave very satisfying results (as the figure shows). In the other 3 forecast points, there were man made structures that were not taken into account by the hydrological model introducing, of course, some bias to the forecast.
INTOGENER has clearly the potential to become a financially self-sustainable flow prediction service for hydropower companies and water management agencies focused on remote mountain areas and unusual years, since it has demonstrated to improve short and mid term forecasts, crucial for energy production optimization.