The European Space Agency (ESA) and the EUMETCast service by EUMETSAT distribute Soil Moisture and Ocean Salinity (SMOS) satellite Near-Real Time (NRT) soil moisture data in less than three hours after sensing. This product is addressed to numerical weather prediction and hydrology operational centers and it is available since 2016. It has been obtained by training a neural network using as input SMOS NRT brightness temperatures for incidence angles in the 30º-45º range and soil temperature from the European Centre for Medium Range Weather Forecast (ECMWF). The SM dataset used as reference to train the neural network was version 620 of the L2 SM product for the NRT SM version 100 dataset. Since August 2018, the new version 200 of the NRT SM dataset is distributed. This new version has been obtained re-training the Neural Network with version 650 of the L2 SM dataset. A preliminary evaluation of the NRT SM v100 dataset was presented in Rodriguez-Fernandez et al. (2017, HESS). This contribution will discuss a more extensive evaluation using more than two years of SMOS NRT SM data. It will also present the first evaluation of the NRT SM version 200 dataset with respect to in situ measurements in a large number of sites from the SMOSMANIA, SCAN, HOBE, USCRN, TERENO, REMEDHUS, PB-H2O and SNOTEL networks. The NRT SM performance will be compared to that of the L2 SM dataset, showing that it is similar but slightly better than that of the L2 SM dataset. Finally, it will be shown that the NRT SM dataset has been able to capture extreme events such as the Harvey floods in Texas in 2017.

Area A - Board 130

The European Space Agency (ESA)’s Soil Moisture and Ocean Salinity (SMOS) mission is the first satellite specifically dedicated to measuring soil moisture (SM) [1]. SMOS was launched in November 2009 and is still in orbit, providing an unprecedented record of L-band brightness temperature (TB) observations at tens of km of spatial resolution (~40 km) with a 3-day revisit time. Since its launch, SMOS has exhibited a successful performance, fulfilling the scientific requirements in terms of both SM and ocean salinity [2]. SMOS as well as the Soil Moisture Active Passive (SMAP) mission have allowed obtaining the most ever accurate SM measurements at global scale [3]. The operational continuity of L-band observations after SMOS and SMAP is now being proposed via the Copernicus Microwave Imaging Radiometer (CIMR) high priority candidate mission [4].
The latest release (v650) of the ESA’s SMOS Level 2 (L2) Soil Moisture User Data Product (SMUDP) has several improvements on the L2 processor [5]. SMOS L2 SMUDP is generated for each orbit at a 15 km Icosahedral Snyder Equal Area (ISEA) 4H9 grid. Due to that, SMOS Level 3 (L3) SM products, which are a composite map of all orbits within the day at a 25 km Equal Area Scalable Earth (EASE)-2 grid, are preferred by research community in some cases to avoid the manipulation difficulties of ISEA. Nevertheless, all SMOS L2 and L3 SM products, and even a higher spatial resolution Level 4 (L4) SM product [6], have demonstrated to be useful for a wide range of scientific and operational applications up to date, being used in predictive hydrological and atmospheric models [7, 8], to monitor flood and drought events [9, 10], to predict wildfire risks [11] and to estimate root zone SM [12].
Nowadays, there are three different available daily SMOS L3 SM products. The first one is a composite of binned data generated by the Centre Aval de Traitement des Données SMOS (CATDS). The SMOS-CATDS L3 SM product is retrieved using a multi-orbit algorithm developed by the Centre d’Etudes Spatiales de la Biosphere (CESBIO) [13]. The second L3 SM product is generated by the Barcelona Expert Centre (BEC). The SMOS-BEC L3 SM product is obtained directly from L2 SMUDP, after applying a filtering based on the Data Quality Index (DQX) parameter and a weighted binning [14]. However, this filtering criterion may be questioned by a general increase of DQX in L2 v650 compared to the previous version (v620) [5]. As an alternative, a filtering based on the retrieval fit quality index, called Chi-Squared (Χi2) parameter, is currently being evaluated. The third L3 SM product is generated by the Institut National de la Recherche Agronomique (INRA) and CESBIO. The SMOS-IC L3 SM product is retrieved with an algorithm that has some simplifications with respect to the official L2 processor, mainly related to the pixel heterogeneity, angle geometry, and vegetation scattering albedo and soil roughness parameters [15].
This study assesses the SMOS-CATDS, SMOS-BEC and SMOS-IC L3 SM products from January 2015 to December 2016. The alternative SMOS-BEC L3 SM filtered by Χi2 (instead of DQX) has also been tested for several thresholds. Different in situ SM networks have been used to validate all possible L3 SM products over several climate types and land covers. Many present and upcoming applications could get benefit of the improvement and refinement of these global SM products. The obtained results and the inferred conclusions will be presented at the conference.

References
[1] Kerr, Y.K.; Waldteufel, P.; Wigneron, J.P.; Delwart, S.; Cabot, F.; Boutin, J.; Escorihuela, M.J.; Font, J.; Reul, N.; Gruhier, C.; et al. (2010) “The SMOS Mission: New Tool for Monitoring Key Elements of the Global Water Cycle”, Proceedings of IEEE, 98: 666-687.
[2] Mecklenburg, S.; Drusch, M.; Kaleschke, L.; Rodriguez-Fernandez, N.; Reul, N.; Kerr, Y.H.; Font, J.; Martin-Neira, M.; Oliva, R.; Daganzo-Eusebio, E.; et al. (2016) “ESA's Soil Moisture and Ocean Salinity mission: From science to operational applications”, Remote Sensing of Environment, 180: 3-18.
[3] Kerr, Y.H.; Al-Yaari, A.; Rodríguez-Fernández, N.; Parrens, M.; Molero, B.; Leroux, D.; Bircher, S.; Mahmoodi, A.; Mialon, A.; Richaume, P.; et al. (2016) “Overview of SMOS performance in terms of global soil moisture monitoring after six years in operation”, Remote Sensing of Environment, 180: 40-63.
[4] Donlon, C.J. (2018) “Copernicus Imaging Microwave Radiometer (CIMR). Mission Requirements Document”, Technical report ESA-EOPSM-CIMR-MRD-3236, revision 1.5, Mission Science Division, European Space Agency (ESA), Noordwijk, Netherlands.
[5] ESA (2017) “Read-me-first note for the release of the SMOS Level 2 Soil Moisture data products”, Technical report, Expert Support Laboratory (ESL) Level 2 Soil Moisture and Array Systems Computing Inc.
[6] Portal, G.; Vall-llossera, M.; Piles, M.; Camps, A.; Chaparro, D.; Pablos, M.; Rossato, L. (2018) “A spatially consistent downscaling approach for SMOS using an adaptive moving window”, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 11(6): 1883-1894..
[7] Ridler, M.E.; Madsen, H.; Stisen, S.; Bircher, S.; Fensholt, R. (2014) “Assimilation of SMOS‐derived soil moisture in a fully integrated hydrological and soil‐vegetation‐atmosphere transfer model in Western Denmark”, Water Resources Research, 50: 8962-8981.
[8] Leroux, D.J.; Pellarin, T.; Vischel, T.; Cohard, J.M.; Gascon, T.; Gibon, F.; Mialon, A.; Galle, S.; Peugeot, C.; Seguis, L. (2016) “Assimilation of SMOS soil moisture into a distributed hydrological model and impacts on the water cycle variables over the Ouémé catchment in Benin”, Hydrological Earth System Sciences, 20: 2827-2840.
[9] Wanders, N.; Karssenberg, D.; de Roo, A.; de Jong, S. M.; Bierkens, M.F.P. (2014) “The suitability of remotely sensed soil moisture for improving operational flood forecasting”, Hydrological Earth System Sciences, 18: 2343-2357.
[10] Pablos, M.; Martínez-Fernández, J.; Sánchez, N.; González-Zamora, Á. (2017) “Temporal and Spatial Comparison of Agricultural Drought Indices from Moderate Resolution Satellite Soil Moisture Data over Northwest Spain”, Remote Sensing, 9:1168.
[11] Chaparro, D.; Vall-llossera, M.; Piles, M.; Camps, A.; Rüdiger, C.; Riera-Tatché, R. (2016) "Predicting the Extent of Wildfires Using Remotely Sensed Soil Moisture and Temperature Trends," IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 9(6): 2818-2829.
[12] Pablos, M.; González-Zamora, Á.; Sánchez, N.; Martínez-Fernández, J. (2018) “Assessment of Root Zone Soil Moisture Estimations from SMAP, SMOS and MODIS Observations”, Remote Sensing, 10: 981.
[13] Al Bitar, A.; Mialon, A.; Kerr, Y.H.; Cabot, F.; Richaume, P.; Jacquette, E.; Quesney, A.; Mahmoodi, A.; Tarot, S.; Parrens, M.; et al. (2017) “The global SMOS Level 3 daily soil moisture and brightness temperature maps”, Earth System Science Data, 9, 293–315.
[14] González-Zamora, Á.; Sánchez, N.; Martínez-Fernández, J.; Gumuzzio, Á.; Piles, M.; Olmedo, E. (2015) “Long-term SMOS soil moisture products: A comprehensive evaluation across scales and methods in the Duero Basin (Spain)”, Physics and Chemistry of the Earth, Parts A/B/C, 83-84: 123-136.
[15] Fernández-Morán, R.; Al-Yaari, A.; Mialon, A.; Mahmoodi, A.; Al Bitar, A.; De Lannoy, G.; Rodríguez-Fernández, N.; López-Baeza, E.; Kerr, Y.H.; Wigneron, J.P. (2017) "SMOS-IC: An Alternative SMOS Soil Moisture and Vegetation Optical Depth Product", Remote Sensing, 9: 457.

Area A - Board 131

The European Space Agency (ESA) launched the SMOS (Soil Moisture Ocean Salinity) mission in 2009; currently, the satellite provides brightness temperature (TB) measurements for multiple global soil moisture (SM) products. The products include the ESA L2SM product; a CESBIO (Centre d'Etudes Spatiales de la BIOsphère) L3SM product, which includes some modifications with respect to the L2 algorithm; a product based on a simplified retrieval algorithm called SMOS-IC (INRA-CESBIO), which is more independent from models and auxiliary data (such as optical vegetation indices), and a near real time product based on a neural network (NN) algorithm. NASA launched the SMAP (Soil Moisture Active Passive) mission in 2015. The SMAP radiometer produces a soil moisture product at a similar spatial resolution and using the same observation frequency as SMOS. In this study, the four SMOS algorithms are compared with each other and with respect to the SMAP radiometer-based SM product. The comparisons are focused over sites with SM networks that include several stations within the SMOS and SMAP footprint area. This dense distribution of stations facilitates upscaling of the pointwise in situ measurements to the SMOS and SMAP footprint scale. The sensors are calibrated to the local soil conditions, and the measurements go through a quality control procedure before being used in the upscaled soil moisture estimate. These sites are called core validation sites (CVS). Thus, the CVS provide an in situ reference that allows the estimation of the performance of each of the SM products at these locations independently. This is a valuable addition to other validation and intercomparison approaches. The study uses the full SMOS lifetime for the SMOS SM product intercomparisons (as the availability of the CVS data allows), and the SMAP lifetime for the comparisons with the SMAP SM product. The quality flags are accounted for each product separately. Performance is assessed between the SM products and the in situ estimates using several metrics. The metrics are computed for all available data, and also separately for days when all products produced a valid retrieval.

Area A - Board 132

Water resources stakeholders across the United States have revealed the need for consistent, high space- and time-resolution, integrated water analyses, predictions, and data to address critical unmet information and service gaps related to floods, droughts, water quality, water availability, and climate change. To meet these growing stakeholder demands and needs, NOAA has undertaken a major effort to improve its hydrological forecast services through the development of a new National Water Model (NWM) at the National Water Center (NWC). The NWM simulates and forecasts land surface runoffs and river discharges using NOAH-MP land surface model estimated land surface soil moisture via response to the precipitation and other forcing data. Currently NWM does not use any external soil moisture information. Because of the uncertainties in model physics and input parameters, and potential errors in forcing data such as precipitation, the soil moisture estimates may be erroneous, resulting uncertainties in the output of the NWM. This study evaluates the satellite soil moisture data products with satellite and in situ soil moisture, identify the data need specifics of the NWM for hydrological predictions. The improvement will use synergistically the observations from VIIRS, AMSR2, ABI, SMAP and SMOS to develop and validate soil moisture data with enhanced spatial resolution as fine as 375m and temporal resolution as short as daily if not hourly in order to meet the potential NWM data requirements. In addition, this study investigates the rainfall impact on soil moisture field variability and contributions toward understanding the impact of the rainfall as a forcing to the soil moisture field. We will present the analysis that attemps to quantify the coupling of soil moisture simulated from the land surface model in NWM and satellite soil moisture with precipitation from GPM and ground radars and gauges, across a range of scale in time and space.

Area A - Board 133

Precipitation measurements are of paramount importance for hydrometeorological applications. Gridded rainfall products aim to provide alternative data sources in regions where gauge measurements are sparse yet their accuracy is to be evaluated. This study examines the performance of NASA's state of the art Integrated Multi-satellitE Retrievals for the Global Precipitation Measurement Mission (IMERG, GPM) satellite precipitation data set in capturing the spatio-temporal variability of weather events compared to the very advanced gauge adjusted German weather radar data set RADOLAN RW. Both data sets aim to be community ready. The comparison setup helps investigating spatial patterns of aggreement between the data sets. Besides quantity also timing of rainfall is of very high importance when modeling or monitoring the hydrologic cycle and therefore a set of detection metrics is evaluated along with standard statistical measures to test both data sets. Additionally, a pixel-by-pixel comparison is performed to assess the ability to represent the spatial variability of rainfall and precipitation quantity. All calculations are additionally carried out for seasonal subsets of the data to address potentially different behavior across precipitation regimes. The results indicate significant temporal and spatial differences between the data sets. Overall, GPM IMERG overestimates the quantity of precipitation compared to RADOLAN, especially in winter season. Moreover, shortcomings in detection performance arise as well in this season with significant erroneously detected yet also missed precipitation events compared to the weather radar data. The differences in the data sets vary over the study region and are clearly spatially distinct, e.g. along secondary mountain ranges and the Alps. Topographically induced precipitation is not well represented in GPM data, which generally shows a lack of spatial variability in rainfall and snowfall estimates due to lower resolution. These regions characterized by high change in elevation on short distance show the highest quantitative differences as well.

Area A - Board 134

Soil water availability greatly affects the growth and condition of vegetation, especially in semi-arid to arid regions. Vegetation condition, in turn, affects fire risk, soil health and can contribute to heatwaves through land-atmosphere feedback processes. Forecasting vegetation condition in response to water availability months ahead would therefore be of great value for timely mitigation of such impacts. At monthly to seasonal time scales, knowledge of water stored in the system, contributes more to predictability than knowledge of the climate system state. However, the lack of knowledge on the vertical distribution of water stores poses the challenge to understand the vegetation and soil water interaction at large scale.
Here we demonstrate that assimilating multiple satellite observations of water status into different soil layers of hydrological models can help to accurately quantify the vertical water distribution at both regional to global scale. SMOS surface soil moisture, MODIS surface water extent, and GRACE total water storage anomalies observations were jointly assimilated into landscape hydrological model using ensemble-based data assimilation methods at both continental (~5 km resolution) and global (~25 km resolution) scales. Significant improvements were found in the estimation of root-zone soil moisture comparing to open-loop model estimations with average correlation against in situ observations increased by 0.1. Analysis estimates of water storage and satellite vegetation observations (NDVI) were used to establish vegetation-accessible water storage relationships. We applied a statistical forecast method using the accessible water storage to forecast NDVI months in advance and compared the results with climatology forecasts. Our findings suggest that incorporating current soil water availability, specific to different vegetation type and aridity, can significantly improve the accuracy of vegetation condition forecasts up to 3 months in advance for the majority of dryland areas of the globe. Such improved forecast lead times have the potential to improve drought early warning system and reduce detrimental impacts on the environment and the economy. This improved capacity may become even more important in the context of projected increases in the occurrence and severity of drought under climate change.

Area A - Board 135

A GEOsynchronous Synthetic Aperture Radar (GEO SAR) mission named Geosynchronous – Continental Land-Atmosphere Sensing System (G-CLASS) has been recently proposed to the European Space Agency (ESA) in response to the call for a future Earth Explore satellite (EE-10). The research presented in this work aims at investigating the potentialities of this very promising satellite mission concept, for hydrological applications. Since a GEO SAR system is able to acquire data with an unprecedented combination of spatial (≤ 1 km) and temporal (≤ 12 h) sampling intervals, this work investigates its potentialities for soil moisture (SM) mapping in order to define the best compromise in terms of image spatio-temporal resolution. To this aim, a synthetic twin soil moisture–data assimilation experiment (SM-DA) was set up to evaluate the impact of the hydrological assimilation of different GEO SAR-like SM products that were simulated with diverse spatial and temporal resolutions.
In the experiment, a ‘‘reference” model simulation (whose outcome is assumed to be the “truth”) was compared against a set of “perturbed” model simulations, where the SM observations were actually assimilated. The “perturbed” model simulations represent a realistic context of hydrological simulations, as such they were obtained by perturbing the precipitation field and the model calibration parameters that were used in the ‘‘reference” model simulation. This was done to simulate some of the main errors that normally affect hydrological modelling. The spatial resolution of all the model simulations was 100 m whereas the temporal resolution was set to 1 h. The hydrological catchment that was used for setting up the experiment was the Cervaro River Catchment, located in the Puglia region (Southern Italy). The reference time period of the experiment was October 2016 - January 2017.
Three GEO SAR-like SM observations were produced by aggregating, both in space and/or in time, the SM maps produced by the ‘‘reference” model simulation: GEO-1 (spatial resolution: 800 m, temporal resolution: 1 h); GEO-2 (spatial resolution: 400 m, temporal resolution: 2 h); GEO-3 (spatial resolution: 100 m, temporal resolution: 8 h). Subsequently, the SM observations were perturbed by adding a white Gaussian noise to simulate the expected instrumental/retrieval error (i.e., 0.07 m3/m3). The simulated maps were then assimilated within the “perturbed” model runs to observe the impact on model discharge predictions and on the modelled SM state estimates.
To understand if GEO SAR-like SM products can provide an added-value with respect to the SM products derived from existing SAR systems, such as Sentinel 1 (S1), the experiment was complemented by an analogous investigation in which the SM observations were simulated as they were produced by a quasi-polar SAR system, like the S1 constellation (POLAR SAR). Three POLAR SAR-like SM observations were produced to have a temporal resolution equal to S1 data (≤ 6 days) and different spatial resolution equal to the GEO SAR ones: POLAR-1 (spatial resolution: 800 m); POLAR-2 (spatial resolution: 400 m); POLAR-3 (spatial resolution: 100 m). The revisit time of the POLAR SAR-like SM products corresponds to that of the S1 A and B images that were actually acquired by the two satellites over the Cervaro River Catchment (relative orbit numbers: 124, 44, 146). After the spatial aggregation, also these three products were perturbed by adding a white Gaussian noise of 0.07 m3/m3. The POLAR SAR-like SM products were finally assimilated within the “perturbed” model runs and results compared against those obtained by the assimilation of the GEO SAR synthetic products.
Findings showed that the assimilation of all GEO SAR-like SM products improved the model performances, both in terms of discharge predictions and in terms of SM state modelling. Amongst the different GEO SAR observations that were assimilated, the product retaining the higher spatial resolution (100 m) and a temporal resolution of 8 h provided the highest improvements on the model performances (i.e., the GEO-3 product). Therefore, for sub-daily observations, as those that can be retrieved from GEO SAR acquisitions, the spatial resolution requirement was shown to be more important that the temporal one. The comparison between the results obtained by assimilating the GEO SAR-like SM products, and the one obtained by assimilating the POLAR SAR-like SM products, showed that the performances of the first one can reach an improvement that is almost 45% higher than the performances of the second one.

Area A - Board 136

Precipitation is one of the key variables in rainfall-runoff modeling. Its reliable estimation is particularly important for mountainous catchments which hydrological response is very fast. Traditionally applied methods for rainfall estimation - rain gauges, have limited applicability over complex terrain due to challenges of its spatial interpolation. Regardless of estimation errors, the weather radars can perform relatively well over mountainous areas, but their spatial coverage is limited. Moreover, radar systems are expensive in construction and difficult to maintain even for developed countries. Launched in 2014 by NASA and JAXA the Global Precipitation Measurement (GPM) mission provides precipitation estimates at quasi-global coverage. The data are available within 4-hour latency as near-real-time, and post-real-time data developed compiled within 2.5 months. One of the developed products under this mission is a gridded rainfall of a 0.1°x0.1° spatial and 30 minutes temporal resolution. The biggest advantages of IMERG GPM data are short time of latency and spatial coverage, which are particularly important for mountainous environments. However, the IMERG GPM advantages are offset by a rather sparse spatial resolution.

We present the assessment of IMERG GPM satellite precipitation estimates for event-based rainfall-runoff modeling in a small mountainous catchment located in southern Poland. The catchment has an area of 240.4 km2 and can be divided into six sub-catchments. In the study, we used a GPM-Level 3 Final Run product which is a post-real time research product released within a latency of 2.5 months. As some of the satellite grids covered the areas outside the sub-catchments (or are common for several sub-catchments), the precipitation weighting was done accounting the area of each sub-catchment covered by each grid. That was done to account for the situation when the rain being observed by the satellite is falling outside of the sub-catchments or partly into several ones. Finally, the hyetographs of 1-hour accumulations were created for each of the sub-catchments and used as precipitation input data.

The hydrological modeling was performed using a semi-distributed HEC-HMS hydrological model run with 1-hour simulation time-step. The simulation results were evaluated using the Nash-Sutcliffe efficiency coefficient (NSE), Pearson’s correlation coefficient (r), percent bias (PBias) and relative peak flow difference (rPFD). The obtained results are encouraging as for the validation phase the NSE ranged from 0.21 to 0.76, r ranged from 0.47 to 0.92, PBias ranged from 3.50 to 14.90% and rPFD ranged from -32.5 to 8.6%. It was proved that low-resolution satellite precipitation estimates can be a valuable source of information on rainfall for small catchments of complex orography and can successfully substitute rain gauges or weather radars. However, it needs to be highlighted that IMERG GPM rainfall estimates seem to be overestimated under extreme rainfall conditions.

Area A - Board 137

Rainfall thresholds are one of the main tools used to characterize and warn against landslide hazard. While main improvements have been made towards more reproducible techniques for the identification of triggering conditions for landsliding, the now well-established rainfall intensity or event – duration thresholds for landsliding suffer from several limitations, e.g., the variable definition of triggering rainfall events and non-consideration of ground conditions. Here, we propose a new approach of the frequentist method for threshold definition based on satellite-derived antecedent rainfall estimates directly coupled with landslide susceptibility data. Adopting a bootstrap statistical technique for the identification of threshold uncertainties at different exceedance probability levels, it results in thresholds expressed as AR = (α ± ∆α)*S^(β ± ∆β), where AR is antecedent rainfall (mm), S is landslide susceptibility, α and β are scaling parameters, and ∆α and ∆β are their uncertainties. In assigning the triggering and causative roles to AR and susceptibility respectively within a renewed 2D frequentist graph, we distribute the influence of hydrological conditions between trigger (by devising an elaborate AR function that aims to reflect the hydrology of an empirical average soil) and cause (by capturing in the susceptibility the spatial variations of hydrological conditions expected from the distribution of the causative ground factors). In this way, we obtain rainfall (AR) thresholds as functions of susceptibility, which enables us to associate threshold mapping with susceptibility maps. Rainfall threshold research is almost inexistent in Africa despite high levels of landslide susceptibility and hazard, especially in mountainous tropical Africa, characterized by intense rainfall, deep weathering profiles and high demographic pressure on the environment. We apply our approach in the western branch of the East African Rift based on landslides that occurred between 2001 and 2018, satellite rainfall estimates from the Tropical Rainfall Measurement Mission Multi-satellite Precipitation Analysis (TMPA 3B42 RT), and the continental-scale map of landslide susceptibility of Broeckx et al. (2018) and present the first regional rainfall thresholds for landsliding in tropical Africa. This work is carried out within the framework of the BELSPO RESIST project.

Area A - Board 138

Early Warning Systems (EWS), generally based on rainfall thresholds, are useful tools for mitigating impacts of shallow landslides and floods. Their necessity is evident, due to the increase in damaging phenomena triggered by intense and concentrated rainfalls caused by climate change. Even if satellite sensors are able to provide soil moisture and rainfall measurements over large areas, threshold based EWS generally do not take into account both parameters. Soil hydrological conditions can play a fundamental role in the triggering of shallow failures or flooding events. Rainfall measures are also uncertain due to limited spatial representativeness of ground sensors and low density of available measuring networks.
This study presents preliminary results of the development of a hydrological model to predict shallow landslides and floods, to define a prototypal EWS useful for Civil Protection and risks mitigation. This system will be validated in three pilot catchments of Oltrepò Pavese area (Northern Italy), which was affected during the last 10 years by numerous events triggered by intense and frequent rainfalls, causing human fatalities, damaging/blocking roads and bridges, destructing cultivations (mainly vineyards). Two integrated hydro-meterological monitoring stations were installed since 2012 in representative geological, geomorphological and environmental settings, in order to identify the conditions leading to rainfall-induced shallow landslide triggering and flood occurrence. Both stations are endowed with Time Domain Reflectometer probes, to measure soil water content and water electrical conductivity, and with a coupled system of tensiometers and dielectric sensors, to measure pore water pressure.
Field measured and satellite derived soil moistures and rainfall amounts are presented and compared. This analysis is fundamental to assess the reliability of these data to be implemented in the integrated EWS.
This research is part of the ANDROMEDA (A New integrateD hydROgeological Model to assEss landsliDes and flood prone Areas in Oltrepò Pavese) project, funded by Cariplo Foundation.

Area A - Board 139

Rainfall is the primary input for hydrologic models that simulate the rainfall-runoff processes at basin scale. Because rainfall is highly variable in space and time, accurate hydrological simulations require accurate rainfall data at the best possible resolution. The conventional rain gauge observations in many parts of the world are sparse and unevenly distributed. Satellite-based rainfall products (SRPs) could be an alternative to traditional rain gauge observations and nowadays are available on a global scale at ever increasing spatial and temporal resolution.
This study proposes a comprehensive assessment of SRPs for flood modeling in Europe. For this purpose, multiple SRPs (i.e., the Tropical Rainfall Measurement Mission (TRMM) Multi-satellite Precipitation Analysis TMPA; the Climate Prediction Center (CPC) Morphing algorithm, CMORPH, the Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks, PERSIANN) will be used to force different lumped hydrologic models (e.g., MISDc, GR4J, HYMOD) over several (+900) basins throughout Europe with different sizes and physiographic characteristics. In particular, this study will allow to: 1) assess the quality of different SRPs for flood modelling and its relationship with climatic/geomorphological conditions; 2) explore the connection between the accuracy of SRPs and their performance in terms of flood modeling taking into account the rainfall-runoff model structure as well. Preliminary results indicated that: 1) satellite rainfall products are not completely reliable for flood forecasting; 2) the hydrological performances of satellite rainfall products depend both on the product and on the selected hydrological model making general guidelines for the optimal use of SRPs in flood modeling difficult to be drawn. To overcome this issue a multimodel-multiproduct approach would help to exploit relative skills of each satellite product-hydrological model configuration and would bring to a more reliable flood forecasting system.

Area A - Board 140