Planning of satellite images applied to early warning hydrological models Estefan´ ıa De El´ ıa, Marcelo Oglietti, Sergio Masuelli, Eduardo Romero CONAE - Argentine National Space Agency CETT, RP C45 Km 8, C´ ordoba Argentina {edeelia}{marcelo.oglietti}{sergio.masuelli}{eromero}@conae.gov.ar Abstract. Space information has become essential for assessing the damages caused by natural emergency situations like earth- quakes, flooding or fires after its occurrence. During the last decade the effort increasingly moved to the objective of us- ing operational systems that combine space information and physical modeling for forecasting and emergency mitigation action planning. These emergency early warning and mitiga- tion support systems require as input space information of a higher quality like high resolution optical or synthetic aper- ture radar (SAR) images which are scarce resources and re- quire an significant latency from request to the actual image acquisition. In this paper we show that for the particular case of flooding it is possible to use a operational medium fidelity physics model to forecast the risk of flooding events and use this result to se- lect in advance which are the most convenient images acqui- sitions to request for the near future. We briefly describe the Reduced Complexity Kinematic Wave Model used to predict potential flood events and the results demonstrate the model ability to replicate the process of runoff behavior over areas with little slope. We use this flood risk map prediction to give higher priority to those observations corresponding with areas of higher risk of flooding, to overcome usual earth ob- servation satellite system reaction time constrains, implying that we can start acquiring useful data from the beginning of the flood event. 1 Introduction During recent years, Earth Observation Satellites (EOS) space missions experienced an important increase both in number and complexity. The main objective behind several EOS missions currently under development it is not anymore just to provide data for the scientific community post fact analysis, but increasingly to develop near real time opera- tional models and applications. Planning and scheduling of several EOS/sensors is an increasingly important problem for space missions because the need of guaranteeing an op- timal use of its resources and a key factor for any near-real time operational application system. This is particularly true for emergency management where currently there are several countries operating and de- veloping space missions with this objective (e.g. Covello et al. (2010) and Wang et al. (2011)). Another interesting example of this operational applications approach for emer- gency is SensorWeb 2.0. Mandl et al. (2008) presents an ambitious space sensor web for disaster management with the objective of facilitate the United States contribution to the Global Earth Observation System of Systems (GEOSS). GEOSS is a worldwide initiative in this direction, with the objective to form a network of EOSs for a wide range of applications in order to provide a real-time picture of the whole planet by sharing all countries sensor resources. This sensor web relays on most important standards in the area like the Open Geospatial Consortium (OGC) and the Sen- sor Web Enablement (SWE) suite. SensorWeb 2.0 intents to present to the user the most simple possible experience inte- grating automatically several space, air and ground sensors, e.g. Moderate Resolution Imaging Spectrometer (MODIS), NASA’s Earth Observing One (EO-1), the US’s Air Force Weather Agency and an Unmanned Aerial System (UAS). The sensor web allows the users to define their regions of interests and then the system automatically detect events of interest. What the users wants to see is automatically ex- ecuted by means of an appropriate workflow and the best available sensors. For example, if a fire is detected by in- specting MODIS data, this automatically triggers a higher resolution instrument like the Hyperion on the EO-1 satel- lite to take a higher resolution image, which in turn also au- tomatically triggers an Unmanned Aerial System take for more detailed imagery. An EOS system response time is all the time needed for an acquisition reception, scheduling, uplink, execution, data downlink, processing and distribution. EOS systems re- sponse time are considered one of the key variables for the success of these applications (Covello et al. (2010)). This is not surprising considering that an EOS system response time usually goes from 24 hours up to a week depending on the capabilities of the mission satellites and ground segment and that for several applications events of interest might have durations of a few hours or days. Being durations of compa- rable order of magnitude implies that even an small change in the response time might imply a difference in the fea- sibility of a particular operational application. Within the