Institut National des Sciences Appliquées de Strasbourg

The word Crowdsourcing –a compound contraction of Crowd and Outsourcing, was used by Howe in order to define outsourcing to the crowd. Beyond cost, benefits for the company can be substantial. It can externalize the risk of failure and it only pays for products or services that meet its expectations. The Crowdsourcing phenomenon covers heterogeneous situations and it has inspired a number of authors. However, we are still lacking a general and synthetic view of this concept. The aim of our work is to characterize Crowdsourcing in its various aspects. First we define of Crowdsourcing, and provide examples that illustrate the diversity of Crowdsourcing practices and we present similarities and differences between Crowdsourcing and established theories (Open Innovation, User Innovation and Open Source Software). Then, we propose and illustrate a typology of Crowdsourcing practices based on two criteria: the integrative or selective nature of the process and the type of tasks that are crowd sourced (simple, complex and creative tasks). In either case, the client firm seeks to mobilize external competencies. Relying upon the crowd can be an adequate method, because of its unique characteristics that are fostered by the Internet. Finally, we present some potential benefits and pitfalls of Crowdsourcing.JEL Codes: O32, L17, L24, L29, M19

The lithium-pilocarpine model of status epilepticus (SE) was used to study the type and distribution of seizure-induced neuronal injury in the rat and its consequences during development. Cell death was evaluated in hematoxylin- and eosin-stained sections and by electron microscopy. Damage to the CA1 neurons was maximal in the 2- and 3-week-old pups and decreased as a function of age. On the other hand, damage to the hilar and CA3 neurons was minimal in the 2-week-old rat pups but reached an adult-like pattern in the 3-week-old animals, and damage to amygdalar neurons increased progressively with age. The 3-week-old animals also demonstrated vulnerability of the dentate granule cells. To evaluate neuronal apoptosis, we used terminal deoxynucleotidyl transferase-mediated biotinylated UTP nick end labeling (TUNEL) stain, confocal fluorescence microscopy of ethidium bromide-stained sections, electron microscopy, and DNA electrophoresis. Neurons displaying all of those features of apoptotic death in response to SE were seen in the CA1 region of the 2-week-old pups and in the hilar border of the dentate granule cells of the 3-week-old animals. Some (3/11) of the animals that underwent SE at 2 weeks of age and most of the animals that underwent SE at 3 or 4 weeks of age (8/11 and 6/8, respectively) developed spontaneous seizures later in life; the latter showed SE-induced synaptic reorganization as demonstrated by Timm methodology. These results provide strong evidence for the vulnerability of the immature brain to seizure-induced damage, which bears features of both necrotic and apoptotic death and contributes to synaptic reorganization and the development of chronic epilepsy.

The growing interest in soft robots comes from the new possibilities offered by these systems to cope with problems that cannot be addressed by robots built from rigid bodies. Many innovative solutions have been developed in recent years to design soft components and systems. They all demonstrate how soft robotics development is closely dependent on advanced manufacturing processes. This review aims at giving an insight on the current state of the art in soft robotics manufacturing. It first puts in light the elementary components that can be used to develop soft actuators, whether they use fluids, shape memory alloys, electro-active polymers or stimuli-responsive materials. Other types of elementary components, such as soft smart structures or soft-rigid hybrid systems, are then presented. The second part of this review deals with the manufacturing methods used to build complete soft structures. It includes molding, with possibly reinforcements and inclusions, additive manufacturing, thin-film manufacturing, shape deposition manufacturing, and bonding. The paper conclusions sums up the pros and cons of the presented techniques, and open to developing topics such as design methods for soft robotics and sensing technologies.

The creation of as-built Building Information Models requires the acquisition of the as-is state of existing buildings. Laser scanners are widely used to achieve this goal since they permit to collect information about object geometry in form of point clouds and provide a large amount of accurate data in a very fast way and with a high level of details. Unfortunately, the scan-to-BIM (Building Information Model) process remains currently largely a manual process which is time consuming and error-prone. In this paper, a semi-automatic approach is presented for the 3D reconstruction of indoors of existing buildings from point clouds. Several segmentations are performed so that point clouds corresponding to grounds, ceilings and walls are extracted. Based on these point clouds, walls and slabs of buildings are reconstructed and described in the IFC format in order to be integrated into BIM software. The assessment of the approach is proposed thanks to two datasets. The evaluation items are the degree of automation, the transferability of the approach and the geometric quality of results of the 3D reconstruction. Additionally, quality indexes are introduced to inspect the results in order to be able to detect potential errors of reconstruction.

The purpose of this study was to compare heart rate (HR) responses within and between physical controlled (short-duration intermittent running) and physical integrated (sided games) training methods in elite soccer players. Ten adult male elite soccer players (age, 26 +/- 2.9 years; body mass, 78.3 +/- 4.4 kg; maximum HR [HRmax], 195.4 +/- 4.9 b x min(-1) and velocity at maximal aerobic speed (MAS), 17.1 +/- 0.8 km x h(-1)) performed different short-duration intermittent runs, e.g., 30-30 (30 seconds of exercise interspersed with 30 seconds of recovery) with active recovery, and 30-30, 15-15, 10-10, and 5-20 seconds with passive recovery, and different sided games (1 versus 1, 2 versus 2, 4 versus 4, 8 versus 8 with and without a goalkeeper, and 10 versus 10). In both training methods, HR was measured and expressed as a mean percentage of HR reserve (%HRres). The %HRres in the 30-30-second intermittent run at 100% MAS with active recovery (at 9 km.h with corresponding distance) was significantly higher than that with passive recovery (85.7% versus 77.2% HRres, respectively, p < 0.001) but also higher than the 1 versus 1 (p < 0.01), 4 versus 4 (p <or= 0.05), 8 versus 8 (p < 0.001), and 10 versus 10 (p < 0.01) small-sided games. The %HRres was 2-fold less homogeneous during the different small-sided games than during the short-duration intermittent running (intersubjects coefficient of variation [CV] = 11.8% versus 5.9%, respectively). During the 8 versus 8 sided game, the presence of goalkeepers induced an approximately 11% increase in %HRres and reduced homogeneity when compared to games without goalkeepers (intersubject CV = 15.6% versus 8.8%). In conclusion, these findings showed that some small-sided games allow the HR to increase to the same level as that in short-duration intermittent running. The sided game method can be used to bring more variety during training, mixing physical, technical, and tactical training approaching the intensity of short-duration intermittent running but with higher intersubject variability.

During long-term fasting in birds and mammals, protein utilization initially decreases (phase I), is thereafter maintained at a low value (phase II), and then further increases (phase III). To delineate hormonal and biochemical changes responsible for these modifications, the effect of food deprivation for 50 days was studied in 6 male king penguins captured at the beginning of their natural breeding fast. During phase II, both rate of mass loss and plasma uric acid concentration remained at low levels, whereas plasma beta-hydroxybutyrate concentration increased. In phase III there was by contrast a 2.5-fold increase in the rate of mass loss, an eightfold increase in plasma uric acid, and an 80% drop in plasma beta-hydroxybutyrate. Plasma corticosterone was low and steady in phase II and increased three times in phase III. During the overall fast, there were no significant variations in plasma insulin, but there was a fourfold increase in plasma glucagon and a decrease in plasma thyroxine and triiodothyronine. These findings suggest that protein sparing (phase II) requires low levels of corticosterone, insulin, and thyroid hormones, whereas the further increase in protein utilization (phase III) is due to an increase in plasma corticosterone. The high plasma glucagon concentration in phase III is presumably responsible for a transient increase in plasma glucose observed at this stage; such increase in glucagon could enhance gluconeogenesis from amino acids.

The goal of this paper is to provide updated diagnostic criteria for the epilepsy syndromes that have a variable age of onset, based on expert consensus of the International League Against Epilepsy Nosology and Definitions Taskforce (2017-2021). We use language consistent with current accepted epilepsy and seizure classifications and incorporate knowledge from advances in genetics, electroencephalography, and imaging. Our aim in delineating the epilepsy syndromes that present at a variable age is to aid diagnosis and to guide investigations for etiology and treatments for these patients.

Abstract The Surface Water and Ocean Topography (SWOT) satellite mission planned for launch in 2020 will map river elevations and inundated area globally for rivers &gt;100 m wide. In advance of this launch, we here evaluated the possibility of estimating discharge in ungauged rivers using synthetic, daily “remote sensing” measurements derived from hydraulic models corrupted with minimal observational errors. Five discharge algorithms were evaluated, as well as the median of the five, for 19 rivers spanning a range of hydraulic and geomorphic conditions. Reliance upon a priori information, and thus applicability to truly ungauged reaches, varied among algorithms: one algorithm employed only global limits on velocity and depth, while the other algorithms relied on globally available prior estimates of discharge. We found at least one algorithm able to estimate instantaneous discharge to within 35% relative root‐mean‐squared error (RRMSE) on 14/16 nonbraided rivers despite out‐of‐bank flows, multichannel planforms, and backwater effects. Moreover, we found RRMSE was often dominated by bias; the median standard deviation of relative residuals across the 16 nonbraided rivers was only 12.5%. SWOT discharge algorithm progress is therefore encouraging, yet future efforts should consider incorporating ancillary data or multialgorithm synergy to improve results.

<para xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> In this paper, an efficient and reliable neural active power filter (APF) to estimate and compensate for harmonic distortions from an AC line is proposed. The proposed filter is completely based on Adaline neural networks which are organized in different independent blocks. We introduce a neural method based on Adalines for the online extraction of the voltage components to recover a balanced and equilibrated voltage system, and three different methods for harmonic filtering. These three methods efficiently separate the fundamental harmonic from the distortion harmonics of the measured currents. According to either the Instantaneous Power Theory or to the Fourier series analysis of the currents, each of these methods are based on a specific decomposition. The original decomposition of the currents or of the powers then allows defining the architecture and the inputs of Adaline neural networks. Different learning schemes are then used to control the inverter to inject elaborated reference currents in the power system. Results obtained by simulation and their real-time validation in experiments are presented to compare the compensation methods. By their learning capabilities, artificial neural networks are able to take into account time-varying parameters, and thus appreciably improve the performance of traditional compensating methods. The effectiveness of the algorithms is demonstrated in their application to harmonics compensation in power systems. </para>

In the last decade, RGB-D cameras - also called range imaging cameras - have known a permanent evolution. Because of their limited cost and their ability to measure distances at a high frame rate, such sensors are especially appreciated for applications in robotics or computer vision. The Kinect v1 (Microsoft) release in November 2010 promoted the use of RGB-D cameras, so that a second version of the sensor arrived on the market in July 2014. Since it is possible to obtain point clouds of an observed scene with a high frequency, one could imagine applying this type of sensors to answer to the need for 3D acquisition. However, due to the technology involved, some questions have to be considered such as, for example, the suitability and accuracy of RGB-D cameras for close range 3D modeling. In that way, the quality of the acquired data represents a major axis. In this paper, the use of a recent Kinect v2 sensor to reconstruct small objects in three dimensions has been investigated. To achieve this goal, a survey of the sensor characteristics as well as a calibration approach are relevant. After an accuracy assessment of the produced models, the benefits and drawbacks of Kinect v2 compared to the first version of the sensor and then to photogrammetry are discussed.

Water spectral indices can enhance the difference between water bodies and background features. Thus, they have been widely used to extract and map surface water bodies based on multispectral satellite imagery. The urban scene is very heterogeneous since the surface is composed of a vast diversity of man-made objects, often of mixed distribution. Urban surface water mapping faces an extreme overestimation phenomenon because certain types of objects such as shadow, dark roads and some artificial features may return similar values to water bodies after an index computation. This study proposes a noise-prediction strategy to eliminate such misclassified nonwater areas in an automated way. Constrained energy minimization (CEM), a typical sparse target detection algorithm that does not need any background information, is utilized to draw the possible distribution of noise based on prior noise samples. The initial noise samples are automatically extracted by calculating the difference between two water indices widely accepted in urban scenes, namely, the modified normalized difference water index (MNDWI) and the automated water extraction index (AWEI). Recently freely available Sentinel-2 multispectral satellite imagery, with high spatial resolution (up to 10 m) and high repeated global coverage (every 5 days), was adopted, considering its potential on urban land cover mapping. Compared with the AWEI based approach, the results show that the proposed noise-prediction approach obtained an improved overall accuracy (increased Kappa coefficient by 0.07 on average), dramatically enhanced user accuracy (by 12.47% on average) with reduced noise, and simultaneously slightly decreased producer accuracy (by −1.19% on average). That is, the proposed method possesses an improvement of the misclassification of nonwater bodies to water bodies and a suppression of the missing of water body extraction at the same time. Finally, the comparative results, with the varying water index segmentation thresholds (−0.2 to 0.3) and an automatic Otsu threshold, indicate the robustness to the threshold of the proposed approach.

Abstract. RGB-D cameras, also known as range imaging cameras, are a recent generation of sensors. As they are suitable for measuring distances to objects at high frame rate, such sensors are increasingly used for 3D acquisitions, and more generally for applications in robotics or computer vision. This kind of sensors became popular especially since the Kinect v1 (Microsoft) arrived on the market in November 2010. In July 2014, Windows has released a new sensor, the Kinect for Windows v2 sensor, based on another technology as its first device. However, due to its initial development for video games, the quality assessment of this new device for 3D modelling represents a major investigation axis. In this paper first experiences with Kinect v2 sensor are related, and the ability of close range 3D modelling is investigated. For this purpose, error sources on output data as well as a calibration approach are presented.

This paper presents an original method, based on artificial neural networks, to reduce the torque ripple in a permanent-magnet nonsinusoidal synchronous motor. Solutions for calculating optimal currents are deduced from geometrical considerations and without a calculation step, which is generally based on the Lagrange optimization. These optimal currents are obtained from two hyperplanes. This paper takes into account the presence of harmonics in the back-EMF and the cogging torque. New control schemes are thus proposed to derive the optimal stator currents giving exactly the desired electromagnetic torque (or speed) and minimizing the ohmic losses. The torque and the speed control scheme both integrate two neural blocks, one dedicated for optimal-current calculation and the other to ensure the generation of these currents via a voltage source inverter. Simulation and experimental results from a laboratory prototype are shown to confirm the validity of the proposed neural approach.

Abstract In this study, rating curves (RCs) were determined by applying satellite altimetry to a poorly gauged basin. This study demonstrates the synergistic application of remote sensing and watershed modeling to capture the dynamics and quantity of flow in the Amazon River Basin, respectively. Three major advancements for estimating basin‐scale patterns in river discharge are described. The first advancement is the preservation of the hydrological meanings of the parameters expressed by Manning's equation to obtain a data set containing the elevations of the river beds throughout the basin. The second advancement is the provision of parameter uncertainties and, therefore, the uncertainties in the rated discharge. The third advancement concerns estimating the discharge while considering backwater effects. We analyzed the Amazon Basin using nearly one thousand series that were obtained from ENVISAT and Jason‐2 altimetry for more than 100 tributaries. Discharge values and related uncertainties were obtained from the rain‐discharge MGB‐IPH model. We used a global optimization algorithm based on the Monte Carlo Markov Chain and Bayesian framework to determine the rating curves. The data were randomly allocated into 80% calibration and 20% validation subsets. A comparison with the validation samples produced a Nash‐Sutcliffe efficiency ( ) of 0.68. When the MGB discharge uncertainties were less than 5%, the value increased to 0.81 (mean). A comparison with the in situ discharge resulted in an value of 0.71 for the validation samples (and 0.77 for calibration). The values at the mouths of the rivers that experienced backwater effects significantly improved when the mean monthly slope was included in the RC. Our RCs were not mission‐dependent, and the value was preserved when applying ENVISAT rating curves to Jason‐2 altimetry at crossovers. The cease‐to‐flow parameter of our RCs provided a good proxy for determining river bed elevation. This proxy was validated against Acoustic Doppler current profiler (ADCP) cross sections with an accuracy of more than 90%. Altimetry measurements are routinely delivered within a few days, and this RC data set provides a simple and cost‐effective tool for predicting discharge throughout the basin in nearly real time.

Chromogranins are a family of acidic soluble proteins which exhibit widespread distribution in endocrine cells and neurons. Chromogranin A (CGA), the major soluble component of the secretory granules in chromaffin cells of the adrenal medulla, is a single polypeptide chain of 431 residues with an apparent molecular mass of 70-75 kDa and a pI of 4.5-5. In mature bovine chromaffin granules about 50% of the CGA has been processed. In the present paper, the structural features of the proteolytic degradation mechanism have been characterized with regard to the possible function of CGA as a prohormone, as suggested by recent studies. CGA-derived components present in chromaffin granules were subjected to either two-dimensional gel electrophoresis or HPLC and the N-terminal of each fragment was sequenced. Immunoblotting with antisera to specific sequences within the CGA molecule were used to characterize these fragments further at their C-terminal. In addition, a similar approach was performed to characterize CGA-derived fragments released into the extracellular space from directly depolarized bovine cultured chromaffin cells. Our results identified several proteolytic cleavage sites involved in CGA degradation. Intragranular processing occurs at 12 cleavage sites along the peptide chain located in both N- and C-terminal moieties of the protein; a preferential proteolytic attack in the C-terminal part was noted. We found that CGA processing also occurs in the extracellular space after release, generating new shorter fragments. The proteolytic cleavage sites identified in this study were compared with the cleavage points which are thought to be involved in generating CGA fragments with specific biological activity: pancreastatin, chromostatin and N-terminal vasostatin fragments. In addition, a new 12-amino-acid CGA-derived peptide corresponding to the sequence 65-76 was identified in the soluble core of purified chromaffin granules. This short peptide was released, together with catecholamines, after stimulation of cultured chromaffin cells suggesting its presence within the storage complex of chromaffin granules. The specific biological activity of this CGA-derived fragment remains to be determined.

Since the first attempts some 20 yr ago in the field of hydrology, random walk (RW) particle tracking as applied to solute transport has experienced profound changes. Concepts and mathematical techniques have improved to the point that numerically difficult problems (e.g., advection‐dominated transport in highly heterogeneous media, or reactive transport) are now much easier to address. Random walk has never been widely used for multiphase flow, probably because numerical dispersion is not a major problem for modeling exercises at large scales. However, vadose zone hydrologic studies often point out very strong variations in fluid velocity over relatively short distances. Random walk methods may be well suited for such studies, a possibility which motivated us to write this review. We first give a comprehensive discussion of the theoretical context of the method. The Fokker–Planck–Kolmogorov equation (FPKE) is established for solute transport, as well as the ordinary Langevin equation and its simplifications for transport of small particles (e.g., colloids). Next, numerical methods are developed for the motion of particles in space. An important section is subsequently dedicated to recent RW concepts in the time domain, and to their application to anomalous (non‐Fickian) transport and inverse problems. Adaptations of RW to transport with solute–solid reactions are also provided, as well as several numerical recipes for resolving a few computational difficulties with the RW method. We purposely did not include any comparisons with Eulerian and Lagrangian approaches. These approaches are discussed at length in several references cited in this review. We note, however, that today's computing capabilities provide new incentives to using RW methods for problems where Eulerian methods are potentially unstable or hampered by numerical diffusion.

Abstract Many processing parameters can be adjusted to optimize the fused filament fabrication (FFF) process, a popular and widely used additive manufacturing techniques for plastic materials. Among those easily adjusted parameters are the nozzle temperature, printing speed, raster orientation, and layer thicknesses. Using poly(ether ether ketone) (PEEK) as the base material, a design of experiments analysis was performed on the main FFF parameters. A response surface methodology was applied to analyze the results and to maximize the output responses. Results have shown that the nozzle temperature is the most influential parameter on tensile properties and the crystallinity degree of printed PEEK by FFF process. Parts produced with optimized FFF parameters were then subjected to an annealing treatment to induce a relaxation of residual stress and to enhance crystallinity. The best properties for 3D printed PEEK parts were achieved with annealed parts prepared at 400°C with a printing speed of 30 mm/s, 0.15 mm layer thickness and raster orientation of [0°/15°/−15°]. The resulting parts have mechanical properties comparable to those of injected PEEK.

OBJECTIVE: Weight loss has been identified as a negative prognostic factor in amyotrophic lateral sclerosis, but there is no evidence regarding whether a high-caloric diet increases survival. Therefore, we sought to evaluate the efficacy of a high-caloric fatty diet (HCFD) for increasing survival. METHODS: A 1:1 randomized, placebo-controlled, parallel-group, double-blinded trial (LIPCAL-ALS study) was conducted between February 2015 and September 2018. Patients were followed up at 3, 6, 9, 12, 15, and 18 months after randomization. The study was performed at 12 sites of the clinical and scientific network of German motor neuron disease centers (ALS/MND-NET). Eligible patients were randomly assigned (1:1) to receive either HCFD (405kcal/day, 100% fat) or placebo in addition to riluzole (100mg/day). The primary endpoint was survival time, defined as time to death or time to study cutoff date. RESULTS: Two hundred one patients (80 female, 121 male, age = 62.4 ± 10.8 years) were included. The confirmatory analysis of the primary outcome survival showed a survival probability of 0.39 (95% confidence interval [CI] = 0.27-0.51) in the placebo group and 0.37 (95% CI = 0.25-0.49) in the HCFD group, both after 28 months (point in time of the last event). The hazard ratio was 0.97, 1-sided 97.5% CI = -∞ to 1.44, p = 0.44. INTERPRETATION: The results provide no evidence for a life-prolonging effect of HCFD for the whole amyotrophic lateral sclerosis population. However, post hoc analysis revealed a significant survival benefit for the subgroup of fast-progressing patients. ANN NEUROL 2020;87:206-216.

Abstract. In active mountain belts with steep terrain, bedrock landsliding is a major erosional agent. In the Himalayas, landsliding is driven by annual hydro-meteorological forcing due to the summer monsoon and by rarer, exceptional events, such as earthquakes. Independent methods yield erosion rate estimates that appear to increase with sampling time, suggesting that rare, high-magnitude erosion events dominate the erosional budget. Nevertheless, until now, neither the contribution of monsoon and earthquakes to landslide erosion nor the proportion of erosion due to rare, giant landslides have been quantified in the Himalayas. We address these challenges by combining and analysing earthquake- and monsoon-induced landslide inventories across different timescales. With time series of 5 m satellite images over four main valleys in central Nepal, we comprehensively mapped landslides caused by the monsoon from 2010 to 2018. We found no clear correlation between monsoon properties and landsliding and a similar mean landsliding rate for all valleys, except in 2015, where the valleys affected by the earthquake featured ∼5–8 times more landsliding than the pre-earthquake mean rate. The long-term size–frequency distribution of monsoon-induced landsliding (MIL) was derived from these inventories and from an inventory of landslides larger than ∼0.1 km2 that occurred between 1972 and 2014. Using a published landslide inventory for the Gorkha 2015 earthquake, we derive the size–frequency distribution for earthquake-induced landsliding (EQIL). These two distributions are dominated by infrequent, large and giant landslides but under-predict an estimated Holocene frequency of giant landslides (&gt; 1 km3) which we derived from a literature compilation. This discrepancy can be resolved when modelling the effect of a full distribution of earthquakes of variable magnitude and when considering that a shallower earthquake may cause larger landslides. In this case, EQIL and MIL contribute about equally to a total long-term erosion of ∼2±0.75 mm yr−1 in agreement with most thermo-chronological data. Independently of the specific total and relative erosion rates, the heavy-tailed size–frequency distribution from MIL and EQIL and the very large maximal landslide size in the Himalayas indicate that mean landslide erosion rates increase with sampling time, as has been observed for independent erosion estimates. Further, we find that the sampling timescale required to adequately capture the frequency of the largest landslides, which is necessary for deriving long-term mean erosion rates, is often much longer than the averaging time of cosmogenic 10Be methods. This observation presents a strong caveat when interpreting spatial or temporal variability in erosion rates from this method. Thus, in areas where a very large, rare landslide contributes heavily to long-term erosion (as the Himalayas), we recommend 10Be sample in catchments with source areas &gt; 10 000 km2 to reduce the method mean bias to below ∼20 % of the long-term erosion.

Abstract. Rainfall-induced landslides are a common and significant source of damages and fatalities worldwide. Still, we have little understanding of the quantity and properties of landsliding that can be expected for a given storm and a given landscape, mostly because we have few inventories of rainfall-induced landslides caused by single storms. Here we present six new comprehensive landslide event inventories coincident with well identified rainfall events. Combining these datasets, with two previously published datasets, we study their statistical properties and their relations to topographic slope distribution and storm properties. Landslide metrics (such as total landsliding, peak landslide density, or landslide distribution area) vary across 2 to 3 orders of magnitude but strongly correlate with the storm total rainfall, varying over almost 2 orders of magnitude for these events. Applying a normalization on the landslide run-out distances increases these correlations and also reveals a positive influence of total rainfall on the proportion of large landslides. The nonlinear scaling of landslide density with total rainfall should be further constrained with additional cases and incorporation of landscape properties such as regolith depth, typical strength or permeability estimates. We also observe that rainfall-induced landslides do not occur preferentially on the steepest slopes of the landscape, contrary to observations from earthquake-induced landslides. This may be due to the preferential failures of larger drainage area patches with intermediate slopes or due to the lower pore-water pressure accumulation in fast-draining steep slopes. The database could be used for further comparison with spatially resolved rainfall estimates and with empirical or mechanistic landslide event modeling.