NH6.4

Raindrop size distributions (DSD) information plays a significant role in many scientific fields, especially in radar meteorology. DSD has spatial and temporal variation across different storm types and climatic regimes. Since the development of polarimetric weather radar, the large-scale DSD estimation has been a long-standing goal in radar meteorology. Traditional polynomial regression algorithms for ground polarimetric radars are widely used to estimate DSD parameters due to their simple methodology and acceptable accuracy. However, a simple polynomial regression may not be able to deeply explore the intrinsic relationship using available observations. This study therefore proposes a DSD retrieval model that uses dual-polarization radar observations based on long short-term memory (LSTM) network techniques. Three schemes of a normalized gamma DSD parameters (LSTM- D₀, LSTM- N_w, and LSTM-µ) are designed with different combinations of polarimetric radar measurement inputs. Results show that all LSTM estimators exhibit better performance than the polynomial regression method. The proposed retrieval model using neural network techniques helps to improve quantitative precipitation estimation of weather radar and make sense of a better understanding of precipitation microphysics.

How to cite: Zhu, J., Chen, E., and Dai, Q.: Raindrop size distribution retrieval model from polarization radar observations using neural network techniques, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-3799, https://doi.org/10.5194/egusphere-egu21-3799, 2021.

The Landslide Hazard Assessment for Situational Awareness system (LHASA) gives a global view of landslide hazard in nearly real time. Currently, it is being upgraded from version 1 to version 2, which entails improvements along several dimensions. These include the incorporation of new predictors, machine learning, and new event-based landslide inventories. As a result, LHASA version 2 substantially improves on the prior performance and introduces a probabilistic element to the global landslide nowcast.

Data from the soil moisture active-passive (SMAP) satellite has been assimilated into a globally consistent data product with a latency less than 3 days, known as SMAP Level 4. In LHASA, these data represent the antecedent conditions prior to landslide-triggering rainfall. In some cases, soil moisture may have accumulated over a period of many months. The model behind SMAP Level 4 also estimates the amount of snow on the ground, which is an important factor in some landslide events. LHASA also incorporates this information as an antecedent condition that modulates the response to rainfall. Slope, lithology, and active faults were also used as predictor variables. These factors can have a strong influence on where landslides initiate. LHASA relies on precipitation estimates from the Global Precipitation Measurement mission to identify the locations where landslides are most probable. The low latency and consistent global coverage of these data make them ideal for real-time applications at continental to global scales. LHASA relies primarily on rainfall from the last 24 hours to spot hazardous sites, which is rescaled by the local 99^th percentile rainfall. However, the multi-day latency of SMAP requires the use of a 2-day antecedent rainfall variable to represent the accumulation of rain between the antecedent soil moisture and current rainfall.

LHASA merges these predictors with XGBoost, a commonly used machine-learning tool, relying on historical landslide inventories to develop the relationship between landslide occurrence and various risk factors. The resulting model relies heavily on current daily rainfall, but other factors also play an important role. LHASA outputs the probability of landslide occurrence on a grid of roughly one kilometer over all continents from 60 North to 60 South latitude. Evaluation over the period 2019-2020 shows that LHASA version 2 doubles the accuracy of the global landslide nowcast without increasing the global false alarm rate.

LHASA also identifies the areas where the human exposure to landslide hazard is most intense. Landslide hazard is divided into 4 levels: minimal, low, moderate, and high. Next, the number of persons and the length of major roads (primary and secondary roads) within each of these areas is calculated for every second-level administrative district (county). These results can be viewed through a web portal hosted at the Goddard Space Flight Center. In addition, users can download daily hazard and exposure data.

LHASA version 2 uses machine learning and satellite data to identify areas of probable landslide hazard within hours of heavy rainfall. Its global maps are significantly more accurate, and it now includes rapid estimates of exposed populations and infrastructure. In addition, a forecast mode will be implemented soon.

How to cite: Stanley, T., Kirschbaum, D., and Emberson, R.: Using satellite soil moisture and rainfall in the Landslide Hazard Assessment for Situational Awareness system, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-5620, https://doi.org/10.5194/egusphere-egu21-5620, 2021.

Rainfall-induced landslides have caused destruction and deaths in South America. Accessing its triggers can help researchers and policymakers to understand the nature of the events and to develop more effective warning systems. In this research, triggering rainfall for rainfall-induced landslides is evaluated. The soil moisture effect is indirectly represented by the antecedent rainfall, which is an input of the ANN model. The area of the Rolante river basin, in Rio Grande do Sul state, Brazil, is chosen for our analysis. On January 5^th, 2017, an extreme rainfall event caused a series of landslides and debris flows in this basin. The landslide scars were mapped using satellite imagery. To calculate the rainfall that triggered the landslides, it was necessary to compute the antecedent rainfall that occurred within the given area. The use of satellite rainfall data is a useful tool, even more so if no gauges are available for the location and time of the rainfall event, which is the case. Remote sensing products that merge the data from in situ stations with satellite rainfall data are increasingly popular. For this research, we employ the data from MERGE (Rozante et al., 2010), that is one of these products, and is focused specifically on Brazilian gauges and territory. For each 12.5x12.5m raster pixel, the rainfall is interpolated to the points and the rainfall volume from the last 24h before the event is accumulated. This is added as training data in our Artificial Neural Network (ANN), along with 11 terrain attributes based on ALOS PALSAR (ASF DAAC, 2015) elevation data and generated by using SAGA GIS. These attributes were presented and analyzed in Lucchese et al. (2020). Sampling follows the procedure suggested in Lucchese et al. (2021, in press). The ANN model is a feedforward neural network with one hidden layer consisting of 20 neurons. The ANN is trained by backpropagation method and cross-validation is used to ensure the correct adjustment of the weights. Metrics are calculated on a separate sample, called verification sample, to avoid bias. After training, and provided with relevant information, the ANN model can estimate the 24h-rainfall thresholds in the region, based on the 2017 event only. The result is a discretized map of rainfall thresholds defined by the execution of the trained ANN. Each pixel of the resulting map should represent the volume of rainfall in 24h necessary to trigger a landslide in that point. As expected, lower thresholds (30 - 60 mm) are located in scarped slopes and the regions where the landslides occurred. However, lowlands and the plateau, which are areas known not to be prone to landslides, show higher rainfall thresholds, although not as high as expected (75 - 95 mm). Mean absolute error for this model is 16.18 mm. The inclusion of more variables and events to the ANN training should favor achieving more reliable outcomes, although, our results are able to show that this methodology has potential to be used for landslide monitoring and prediction.

How to cite: Lucchese, L. V., de Oliveira, G. G., and Pedrollo, O. C.: Towards the effective use of Artificial Neural Networks for accessing rainfall thresholds for rainfall-induced landslides, a study based on in-situ and satellite merged rainfall data, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1623, https://doi.org/10.5194/egusphere-egu21-1623, 2021.

Rainfall-induced shallow landslides affect buildings, roads, facilities, cultivations, causing several damages and, sometimes, loss of human lives. It is necessary assessing the most prone zones in a territory where these phenomena could occur and the triggering conditions of these events, which generally correspond to intense and concentrated rainfalls. The most adopted methodologies for the determination of the spatial and temporal probability of occurrence are physically-based models, that quantify the hydrological and the mechanical responses of the slopes according to particular rainfall scenarios. Whereas, they are limited to be applied in a reliable way in little catchments, where geotechnical and hydrological characteristics of the materials are homogeneous. Data-driven models could constraints these, when the predisposing factors of shallow instabilities, allowing to estimate only the susceptibility of a territory, are combined with triggering factors of shallow landslides to allow these methods to estimate also the probability of occurrence and, then, the hazard. This work presents the implementation of a data-driven model able to assses the spatio-temporal probability of occurrence of shallow landslides in large areas by means of a data-driven techniques. The models are based on Multivariate Adaptive Regression Technique (MARS), that links geomorphological, hydrological, geological and land use predisposing factors to triggering factors of shallow failures. These triggering factors correspond to soil saturation degree and rainfall amounts, which are available thanks to satellite measures (ASCAT and GPM). The methodological approach is testing in different catchments of Oltrepò Pavese hilly area (northern Italy), that is representative of Italian Apeninnes environment. This work was made in the frame of the project ANDROMEDA, funded by Fondazione Cariplo.

How to cite: Vivaldi, V., Bordoni, M., Brocca, L., Ciabatta, L., and Meisina, C.: Using satellite soil moisture and rainfall data for the monitoring and the prediction of natural hazards, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14812, https://doi.org/10.5194/egusphere-egu21-14812, 2021.