Scenarios of Drought Impact on River Discharge

Introduction

Figure 1. Dryness in Lake Klöntal, Switzerland. Credits: Sarah Loetscher

Droughts are one of the main water-related geophysical hazards in Europe¹. An area of particular concern is the Mediterranean region due to an increase in the occurrence of these events². While the impact of droughts on food security is acknowledged, its effects span the entire economy at different levels³.

Precipitation and temperature are the main factors that influence the occurrence of droughts. On the one hand, liquid precipitation or rainfall (hereinafter referred to as precipitation) contributes to the river discharge over a short time-span. On the other hand, rising temperatures lead to increasing evapotranspiration, culminating in the reduction of the available water.

Snow is a crucial variable in mountain environments. In the Alps, a substantial proportion of winter and early spring precipitation falls as snow above an altitude of 1,000 meters, which remains stored in the snowpack, and it is slowly released during the warmer months supplying water from April to September. Consequently, the accumulation of snow can alleviate or exacerbate the effects of droughts.

Understanding the behaviour of snow helps to anticipate whether the current season will either experience a shortage, average conditions, or an abundant water availability. This knowledge informs decision-making processes, facilitating the adequate water management and implementation of preventive actions to optimize water availability. As a result, snow cover estimation is one of the main indicators to predict possible drought conditions⁴, also playing a pivotal role in operational flood forecasting⁵.

Droughts can also be caused by human activities in the river basins, such as hydroelectric production, civil water distribution, and irrigation. These activities causing water extraction or restitution, disrupt the natural flow of water within a catchment. Consequently, such operations decrease or increase water availability at specific areas within the basin.

In the following pages, a river discharge and drought tool for decision-making is presented. The selected area to show-case the capabilities of the tool is the Sarca river basin in Italy. The data used in the pipelines of the models that feed into the platform correspond to the period 2002 – 2022, enabling the user to run scenarios of water availability for a given month between April and September based on total averages over the reference period (2002 - 2022).

---

¹ Markonis, Y., Kumar, R., Hanel, M., Rakovec, O., Máca, P., & AghaKouchak, A. (2021). The rise of compound warm-season droughts in Europe. Science advances, 7(6), eabb9668.
² M. Hoerling, J. Eischeid, J. Perlwitz, X. Quan, T. Zhang, P. Pegion, On the increased frequency of mediterranean drought. J. Climate 25, 2146–2161 (2012).
³ K. Stahl, I. Kohn, V. Blauhut, J. Urquijo, L. De Stefano, V. Acácio, S. Dias, J. H. Stagge, L. M. Tallaksen, E. Kampragou, A. F. Van Loon, L. J. Barker, L. A. Melsen, C. Bifulco, D. Musolino, A. de Carli, A. Massarutto, D. Assimacopoulos, H. A. J. Van Lanen, Impacts of European drought events: insights from an international database of text-based reports. Nat. Hazards Earth Syst. Sci. 16, 801–819 (2016).
⁴ Staudinger, M., Stahl, K., and Seibert, J. (2014), A drought index accounting for snow, Water Resour. Res., 50, 7861– 7872, doi:10.1002/2013WR015143.
⁵ Rössler, O., Froidevaux, P., Börst, U., Rickli, R., Martius, O., and Weingartner, R.: Retrospective analysis of a non forecasted rain-on-snow flood in the Alps – a matter of model limitations or unpredictable nature?, Hydrol. Earth Syst. Sci., 18, 2265–2285, https://doi.org/10.5194/hess-18-2265-2014, 2014.

A Decision Support System for Droughts in the Alpine Environment

Figure 2. Mountain basin in the alps. Credits: Sergio Cerrato

The generation of water availability scenarios allows the understanding of how certain changes in the environment could affect the available water in a specific area. The tool can be used to estimate potential damage, or it can also be applied retrospectively for territorial planning, being a valuable tool for decision-makers. There’s also room for research and development on top of this tool, advancing scientific understanding of drought and its impacts.

Through the Decision Support System (DSS) of the DTA demonstrator platform, the user can consult the available water during the spring and summer months (i.e., April until September) up to a sub-catchment scale (i.e., on each river segment) in the Sarca river basin, Italy. The platform enables the user to run scenarios based on the reference period 2002 – 2022, interactively selecting from low, average, and high conditions for each of the variables (i.e., precipitation, temperature, and snow storage). As for human intervention, it can be set to active, considering human activities along the river or inactive, representing the river as it would be in a natural state see Table 1.

When displaying pre-defined scenario layers in the demonstrator, a combination of low, average, or high conditions of the variables that influence river discharge are applied. These conditions are statistically defined based on percentiles calculated over the reference period 2002-2022. Advanced users who would like more control over the simulation process can tailor the tool further and generate their own combination of input variables through a graphical interface or a code box. This process requires nevertheless an understanding of what each percentile represents. To support the customisation, Table 1 presents all the information needed on percentiles, and Page 5 provides a step-by-step explanation for running the river discharge and drought analysis in the DTA.

Once the variables are set, the river discharge and drought DSS displays maps and time series of water availability scenarios across the river basin. Water availability is classified as very low, low, average, high, and very high. The legend of the maps is provided in Table 2.

​​​​Precipitation and temperature were calculated based on historical data taken from the ERA5 dataset (ECMWF), corresponding to the period 2002 – 2022. Snow storage was determined with the GEOtop6 model for the same​ reference period.​

Variable	Description	Percentiles7	Alias of the variable used to create new scenarios8
Precipitation	Provides the water input into the system.	20: low precipitation 50: average precipitation 80: high precipitation	PQ
Temperature	Regulates snowmelt processes, variations of the potential evapotranspiration and separation between liquid and solid precipitation.	20: low temperature 50: average temperature 80: high temperature	TQ
Snow storage	Determines the water storage for the spring and summer period.	20: low snow storage 50: average snow storage 80: high snow storage	SQ
Human intervention	Remove or add water mass into the system.	0: all intake works are disabled or not active, and the reservoirs release the same amount of water as they receive as input, thus simulating a completely natural catchment. 50: all intake works are active. Simulation where water withdrawals at intake or reservoir points are considered.	A

Table 1. Natural and human variables considered in the simulation of scenarios.

Category	Symbol
Very low water availability
Low water availability
Average conditions
High water availability
Very high water availability

---

⁶ The aliases indicated in the table are shown in custom script.
⁷ The values for precipitation, temperature, and snow storage are provided in percentiles, whereas, anthropic works can only be set set to inactive (0) or active (50).
⁸Endrizzi, S., Gruber, S., Dall'Amico, M., & Rigon, R. (2014). GEOtop 2.0: simulating the combined energy and water balance at and below the land surface accounting for soil freezing, snow cover and terrain effects. Geoscientific Model Development, 7(6), 2831-2857.

Examples using the DSS to evaluate drought conditions in the alpine environment

Case 1 – Human intervention effect

Figure 3. Roselen Dam, France. Credits: Freddy (dendoktoor)

In this use case, an evaluation of the effect of human intervention was conducted in the Sarca river basin, Italy. The natural inputs (i.e., precipitation, temperature, and snow storage) were fixed to average conditions, and human intervention was set to inactive to analyse the scenario without human intervention, and to active to analyse the scenario with human intervention (see Table 1). Both simulations correspond to July for the reference period (2002 - 2022).

In a scenario without human activity, the main channel of the river would have high water (Figure 4A), whereas with human intervention, the main channel is negatively impacted with low and very low water availability. The magnitude of the effect is higher in the lower part of the basin, where water extraction to sustain economical activities results in a very low water supply (Figure 4B).

With the DSS tool from the DTA platform, we were able to quickly identify vulnerable sections of the river that already experience reduced water availability in July without considering other changes in the environment such as less rainfall, less snowfall, and/or higher temperature.

Figure 4. Human impact effect over the Sarca river basin, Italy in July. A: without human impact, B: with human impact (See Table 2 to consult the legend of the map). Blue pins represent the location of river discharge measurement stations. Yellow pin indicates a severely affected area of interest (see Figure 5 for time series).

A time series of the river segment which is more dramatically affected (Figure 4, yellow pin) shows that in the period April – September, water availability remains around the average or increases without human intervention. However, with human intervention, after starting with average conditions at the beginning of the spring season, declines up to the 10th percentile (very low water availability) from June to August (Figure 5). Consult Table 1 for the explanation about percentiles.

Case 2 – Snow storage effect

Figure 6. Grossglockner mountain, Austria. Credits: Julius Silver

The contribution of snow to water availability over a river basin can be significant, especially in mountain areas such as the Alps, during spring and summer. Thus, the effect of snow storage on river discharge can be evaluated by varying the snow storage levels and keeping the other inputs fixed.

The scenarios reported in Figures 7A and 7B were taken from August and consider the following common conditions:

• Low precipitation based on the historical mean;
• High temperature based on the historical mean;
• Human intervention is set to active.

Both scenarios, with high (Figure 7A) and low snow storage (Figure 7B) present low and very low water availability to a lesser or greater extent. In the scenario with low snow storage (Figure 7B), the basin is prompt to experience more critical conditions, also affecting the higher altitudes of the basin.

The yellow pin placed in Figure 7A, corresponds to a sector in the upper basin that switches from average conditions (Figure 7A) to low water availability (Figure 7B). On the other hand, the yellow pin placed in Figure 7B, represents a sector in the lower basin where the impact is equally significant and water availability remains very low.

Figure 7. Snow storage effect over the Sarca river basin, Italy in August. A: high snow storage, B: low snow storage. Blue pins represent the location of river discharge measurement stations. Yellow pins indicate the location of river sectors in the A: upper basin and B: lower basin (see Figure 8 for time series).

To further explore the effect of snow storage in the two selected points detailed above, a time series is presented in Figure 8. Overall, the lower basin seems not to be affected by an increase or decrease of snow storage and the other variables such as lower precipitation, higher temperature or human impact might have a higher influence. However, in the upper basin, lower snow storage negatively impacts water availability from July to September.

Figure 8. Time series of water availability under scenarios with low and high snow storage in the lower and upper basin. To generate your own graphs, see the next page.

How to do your own river discharge and drought analysis in the DTA?

1. To gain access to the DTA demonstrator platform, sign up for free using this link.
2. After accessing the demonstrator, select the Sarca river basin, Italy. River discharge and drought scenarios are provided for this region.

3. The blue pins in the map represent the location of river discharge measurement stations in the Sarca river basin. Next, click on any of the pins to get river discharge data measured at the selected station for the period 2017 – 2022 or click on the River discharge icon located on the left panel.

4. A time series chart of river discharge data is presented. The box next to the selected station is activated. Modify the Time range and expand or reduce the time series to your needs. Then, click in the Reload icon next to the Time range.

5. Compare river discharge data measured at the Ponte Arche station, with other stations in the basin by activating the corresponding boxes. To see the location of the stations, close the time series visualizer to go back to the map. Export the values using the Export CSV button.

6. Now, let’s go a step forward simulating scenarios. Under the water resources management sub-theme, select the Drought Monitoring information layer followed by the Drought DSS option. Modify the Time range accordingly, the anomalies were calculated based on the reference period (2002 - 2022), with one value available for each month from April to September, for the entire period. The layers are available in 2022. Set the starting date on (or before) April 1^st and the end date on (or after) September 1^st of 2022. Then, click Search.

7. The search results return a list of available tiles, select a tile to visualize it.

8. Predefined scenarios are already prepared and presented in the left panel. For instance, in this example we observe a scenario with higher precipitation in September. The different widths observed in the sections of the river, indicate the contributing area of each river segment to the river basin. Refer to Table 2 to consult the legend.

9. In addition to predefined scenarios, it is also possible to create your own scenario and visualize the results on the map. For that, navigate to the end of the left panel, click on Custom and select Custom script to display a code editor where the script can be customized.

10. Creating a custom scenario involves customizing the script let band = "PQxxTQxxSQxxAxx" with the desired values for each condition. In the following example, we want to create a scenario to analyse the effect of lower precipitation and higher temperature, assuming that there are no human water intake structures in the basin. For that purpose, the following considerations are taken, see Table 1 for further details:
• Precipitation (PQ): set to low precipitation (percentile 20)
• Temperature (TQ): set to high temperature (percentile 80)
• Snow storage (SQ): set to average conditions (percentile 50)
• Anthropic works (A): set to no activity (0)

Modify the code to the following: let band = "PQ20TQ80SQ50A0. Then, click on Refresh Evalscript. In case human impact wants to be considered, change A value from 0 to 50. A new visualization representing the custom scenario will be displayed on the map.

11. Note that the scenario obtained in the previous step corresponds to September. To visualize the maps of the other months, move to the top of the left panel and change the Date. Use the left and right arrows to navigate through the different months. The image below corresponds to May.

12. The time series values for the 6-month period April – September can also be obtained for the created custom scenario. To get the full time series, select the desired end Date to September, then select the point of interest on the map using the Mark point of interest icon on the right panel. Next, click on the Drought DSS icon, to retrieve the time series results.

13. By default, the chart displays a time series of the past 6 months, where all the variables are set to average conditions, and human intervention to active. Adjust the input variables to low, average, high, active or not active. Then, click on Add layer. In this example the following conditions were applied:
• Precipitation: low
• Temperature: high
• Snow storage: normal
• Human intervention: not active

14. As observed in the previous step, for the selected point, additional scenarios can be added and be simultaneously compared. For instance, to create a similar scenario considering human impact, modify human intervention from Not active to Active. Next, click on +Add layer. An additional item is added to the list of Available layers, and the corresponding visualization of the time series is included in the chart. The layers are very descriptive of the scenario that they represent, indicating the variable and status of the variable (e.g., Temp-Low, Snow-High, etc.). More scenarios can be added and compared. The values can also be exported as CSV by clicking on Export CSV.

15. After exporting the values as CSV, a chart can be plotted either in excel or using any programming language. The values obtained from different river segments can also be compared in this way, obtaining charts as the one presented in Figure 8.

16. Explore more tools to download images, measure distances, compare maps that correspond to different dates or scenarios, and share your maps in blogs. Find a video tutorial in our website to navigate through the additional functionalities of the demonstrator platform.

Try it yourself following this link

Running Scenarios of Flood impact in the Lower Basin of the Sarca River, Italy

Impact of the flood in the Sarca’s lower basin (Prato Saiano, Arco town in Trento, October 2020).
The picture shows inundated agricultural lands, sport infrastructure and green areas. Credits: Fire department of Trento Province.

Alpine environments are notably susceptible to flooding, due to their intricate topography, abundant river networks, and the influence of snowmelt. The impact of such floods can be exacerbated by factors such as scarce vegetation cover and steep slopes of the terrain¹.

An illustrative example of this hazard and its impacts occurred in the Sarca river basin, Italy, in October 2020. During this event, the region witnessed a severe flood event that damaged the infrastructure and posed a significant threat to local population^2,3. On October 3^rd, within a matter of hours, the water level increased dramatically from 1 m to 5.41 m in a few hours⁴. River discharge also increased from 30 – 50 m/s to 300 m/s, affecting the towns of Arco, Ponte Arche and Dro⁵ (Figure 1).

The occurrence of floods is primarily determined by precipitation (mainly rainfall) and soil moisture. These factors contribute to the available water budget within a basin. In the Alpine region, the amount of rainfall and its intensity is defined by the freezing level and return period (given in years). In the next page, Table 1 provides further details about the main variables that influence floods.

In the following pages, a flood tool for decision making is presented, enabling the simulation of flood scenarios. The area showcased for this use-case is the lower basin of the Sarca river in Italy.

---

¹ Alderman, K., Turner, L. R., & Tong, S. (2012). Floods and human health: a systematic review. Environment international, 47, 37-47.
² https://www.ladige.it/cronaca/2020/10/04/il-sarca-esonda-ad-arco-e-fa-paura-anche-oggi-sorvegliato-speciale-1.2532437
³ https://www.ildolomiti.it/cronaca/2020/foto-e-video-maltempo-il-fiume-sarca-esonda-in-diversi-punti-chiuso-il-ponte-di-arco-e-si-attende-londa-di-piena
⁴ https://www.ladige.it/territori/riva-arco/2020/10/05/paura-e-danni-sul-sarca-esondazioni-e-allagamenti-accuse-alla-diga-di-ponte-pia-1.2533164
⁵ https://www.ildolomiti.it/cronaca/2020/foto-e-video-maltempo-il-fiume-sarca-esonda-in-diversi-punti-chiuso-il-ponte-di-arco-e-si-attende-londa-di-piena

A Decision Support System (DSS) for floods in the alpine environment

Figure 2. A flood event and impacts on the infrastructure and public spaces. Source: https://pixabay.com/users/hans-2/

The generation of flood scenarios allows the understanding of how certain environmental variables influence the occurrence and magnitude of floods. The tool can be used to estimate potential damage, or it can also be applied retrospectively for territorial planning, being a valuable tool for decision-makers. This tool can also be used as a basis for research and development, advancing scientific understanding of floods and its impacts.

Through the Decision Support System (DSS) of the DTA demonstrator platform, you can consult the potential extent of a flood event in the lower basin of the Sarca river in Italy. The tool enables an interactive selection of input conditions for each of the variables (i.e., return period, freezing level, and soil moisture), and displays the peak discharge together with the potential extent and depth of the flood that results from the selected inputs. A description of each of the variables and input conditions can be consulted in Table 1.

Variable [units]	Description	Input conditions
Return period [years]	Describes the average time between intense precipitation events of the same magnitude	10: short term return period 30: medium term return period 100: long term return period
Freezing level [m.a.s.l]	Determines whether precipitation mainly falls as rain or snow.	1000: precipitation primarily falls as snow 2000: solid and liquid precipitation almost fall in the same proportion. 3000: precipitation primarily falls as rain
Soil moisture	Describes the soil absorption capacity to retain water	wet: lower absorption capacity, increases the likelihood of flash floods. dry: higher absorption capacity, reduces the likelihood of floods

Table 1. Variables considered in the simulation of flood scenarios.

Examples of potential flood evaluation using the DSS in an alpine environment

Figure 3. Alpine landscape where the freezing level is clearly outlined. Above the freezing level, is the zone with snow accumulation where precipitation falls mainly as snow. Below the freezing level, is the zone where precipitation predominantly falls as rain.
Source: Maxim Lamare

Case 1 - Freezing level effect

In this use case, an evaluation of the effect of the freezing level was conducted in the lower basin of the Sarca river in Italy, where an intense flood occurred in 2020. Figures 4A and 4B report scenarios of low (i.e., 1000 m.a.s.l), and high freezing levels (i.e., 3000 m.a.s.l), respectively. Both scenarios share common input conditions for return period and soil moisture, which were defined as follows:

• Medium term return period: 30 years
• Soil moisture: wet

A remarkable difference is observed among both scenarios. In a low freezing level scenario, the likelihood of flood occurrence is non-existent (Figure 4A). On the other hand, the extent of a flood in a high freezing level scenario is large (Figure 4B). In the alpine region, the freezing level determines the amount of rainfall and snowfall. In Figure 4A, due to the low freezing level, precipitation is primarily snow. The contribution of snowmelt to rivers and streams is slower than rainfall. On the other hand, Figure 4B presents a scenario where precipitation is primarily rainfall, which rapidly recharges rivers and streams, ultimately leading to floods.

With the DSS tool, the user can quickly assess exposed areas to floods, and obtain an insight of the magnitude of the impact, identifying the areas that may be more vulnerable based on the selected conditions.

Figure 4. Freezing level effect on flood occurrence over the lower basin of the Sarca river, Italy. A: low freezing level, B: high freezing level. The legend represents the water column depth of inundated areas.

Case 2 - Soil moisture effect

Figure 5. Soil moisture condition is a fundamental variable for the assessment of potential flood occurrence. Source: Philippe Bally

Soil moisture plays a relevant role, being a key factor to determine if a flood is likely to happen or what the magnitude of the event would be. When the soil is completely saturated with water, its capacity to absorb more water is limited. Consequently, after a heavy rainfall, excess water leads to increased runoff and a higher likelihood of flooding. The effect of soil moisture was assessed in the platform by considering a scenario with dry (Figure 6A) and wet (Figure 6B) soil moisture. The scenarios also consider the following common conditions:

• Short term return period: 10 years
• Freezing level: 3000 m.a.s.l

Both scenarios indicate that under the selected conditions, the lower basin of the Sarca river would be flooded to a lower (Figure 6A) or greater (Figure 6B) extent. Although both maps are not remarkably different, the DSS shows that (1) soil moisture does influence flood occurrence and/or its magnitude, and (2) the compound effect of the other factors (i.e., return period and freezing level) should be also considered when evaluating the extent of the impact.

Figure 6. Soil moisture effect on flood occurrence over the lower basin of the Sarca river, Italy. A: dry, B: wet. The legend represents the water column depth of inundated areas.

How to do your own river discharge and drought analysis in the DTA?

1. To gain access to the DTA demonstrator platform, sign up for free using this link.
2. After accessing the demonstrator, select the Sarca river basin, Italy. River discharge and drought scenarios are provided for this region.

3. You will notice that the Flood DSS icon on the right of the Browser panel is deactivated (greyed out). In order to activate it, click on the Compare tab in the left panel.

4. Now you can click on the Flood DSS icon. A new window is displayed providing an interface with options to adjust the input conditions (i.e., return period, freezing level, and soil moisture), create flood scenarios, and interactively visualise a discharge - precipitation hydrograph.

5. Once the input conditions are adjusted, click on Add to compare and the map that results from the scenario is displayed in the background. Note that the layer is added within the Compare tab. Run multiple scenarios modifying the input conditions and click on Add to compare to visualise and compare the maps. Every time a parameter is modified, the hydrograph changes. Additionally, when a new scenario is added to the map, it will appear at the top of the compare tool.

6. Close the Flood DSS window and proceed to inspect the maps. To compare scenarios, use the Split and Opacity sliders. Both sliders enable the visualisation and comparison between scenarios. With these sliders, you can also display the underlying basemaps (satellite image or map), to obtain a grasp of the land use and infrastructure more likely to be affected under a specific scenario.

7. Proceed to explore more tools to measure distances and download images . Follow a video tutorial on the website to navigate through the additional functionalities of the demonstrator platform.

Try it yourself following this link

Monitoring Large-scale Landslides Using Sentinel-2 Derived Products

Figure 1. Gran Sasso massif, Italy. Source: Philippe Bally

Deep-seated landslides are a type of large-scale landslide, characterised by their significant depth and substantial displacement of rock, soil, and debris. They are commonly classified as slow-moving landslides, typically at average velocities ranging from a few millimetres to a few decimetres per year. However, they can suddenly accelerate and trigger fast-moving propagation processes, with devastating impacts on infrastructure, ecosystems, and human livelihoods, sometimes posing a risk to human lives.

Large landslides are influenced by continuous deformation of the Earth surface, such as land uplift due to glacier melting, soil swelling resulting from groundwater recharge, subsidence or sinking of the Earth’s surface, among other factors. Precipitation in the form of rain and snow plays a fundamental role regulating these processes by infiltrating into and around the unstable areas and also through deep-water circulation.

Deeper and more extensive failure surfaces in large landslides delay the effect of water infiltration, affecting both acceleration and deceleration phases. This delay is challenging to assess, especially when monitoring data for unstable slopes is scarce or insufficient. Fortunately, access to large amounts of freely available satellite data, particularly the Copernicus Sentinel-1 and Sentinel-2 imagery, along with the development of innovative processing approaches, allows the usage of Earth Observation data to document landslide motion from space.

In the following pages, the deep-seated landslides dataset is presented. The specific areas showcased are La Valette (France) and La Barmasse (Switzerland).

Description of the deep-seated landslides products

Figure 2. Landslide impact on a forest ecosytem. Source: https://pixabay.com/de/users/hans-2/

Monitoring and documenting deep-seated landslides is vital for public safety, minimising adverse impacts on the infrastructure and the environment, and empowering communities to prepare for and mitigate potential economic losses. Continuous monitoring also contributes to the advancement of scientific knowledge, essential to understand and adapt to changing conditions, including climate change.

The deep-seated landslide products disseminated through the DTA platform are based on image correlation algorithms applied to optical Sentinel-2 imagery, processed in the GDM-OPT-SLIDE^1,2 service accessible on the Geohazards Exploitation Platform (GEP). Sentinel-2 archives over the period 2015 - 2023 were processed, providing valuable insights into the deep-seated landslides in the La Valette, France^3,4 and the La Barmasse, Switzerland^5,6.

The deep-seated landslide products are available at a spatial resolution of 10 metres, and it includes the following:

1. Average horizontal velocity map over the period 2015 – 2023
2. Average horizontal East - West displacement map
3. Average horizontal North - South displacement map
4. Time series East - West displacement maps
5. Time series North - South displacement maps
6. Quality indicator: Root Mean Square (RMS) time series of East - West displacement maps
7. Quality indicator: Root Mean Square (RMS) time series of North - South displacement maps

---

¹ Stumpf, A., Michéa, D., Malet, J.-P. 2018. High-precision co-registration of Sentinel-2 and Landsat-8 imagery for surface motion measurements. Remote Sensing, 10(2): 160. doi:10.3390/rs10020160
² Provost, F., Michéa, D., Malet, J.-P., Stumpf, A., Doin, M.-P., Lacroix, P., Boissier, E., Pointal, E., Pacini, F., Proy, C., Bally, P. 2022. Terrain deformation measurements from optical satellite imagery : The MPIC-OPT processing services for geohazards monitoring. Remote Sensing of Environment, 274, 112949. doi:10.1016/j.rse.2022.112949
³ Raucoules, D., de Michele, M., Malet, J.-P., Ulrich, P. 2013. Time-variable 3D ground displacements from High-Resolution Synthetic Aperture Radar (SAR). Application to La Valette landslide (South French Alps). Remote Sensing of Environment, 139: 198-204. doi:10.1016/j.rse.2013.08.006
⁴ Travelletti, J., Malet, J.-P., Samyn, K., Grandjean, G., Jaboyedoff, M. 2013. Control of landslide retrogression by discontinuities: evidences by the integration of airborne and ground-based geophysical information. Landslides, 10: 37-54. doi:10.1007/s10346-011-0310-8
⁵ Michoud C., Abellán A., Baillifard F., Derron M.-H., Jaboyedoff M. and May-Delasoie F.: Using active terrestrial remote sensing techniques to monitor and model the structurally-controlled rockslide of La Barmasse (Valais, Switzerland). Innovative LiDAR Solutions Conferences, Nice, France, 2012.
⁶ Abellán, A., Michoud, C., Jaboyedoff, M., Baillifard, F., Demierre, J. and Carrea, D., 2013, April. Landslide velocity prediction using a rainfall to displacements transfer function. La Barmasse case study (Valais, Switzerland). In EGU General Assembly Conference Abstracts (pp. EGU2013-11813).

Examples of use cases with data from the DTA platform

Case 1 - Monitoring landslide velocity and displacement in the La Barmasse landslide

Figure 3. The La Barmasse landslide, Switzerland. Source: Clément Michoud

In this use-case, an assessment of the velocity and the direction of the displacement was conducted in La Barmasse, a northeast facing unstable slope situated in the Val de Bagnes valley in the Valais canton, Switzerland. La Barmasse slope is an active area for rock and debris slides. It threatens the Dranse river and a major road on which significant cracks have been noted since 2013. This slope is continuously and slowly moving with rates controlled by intense rainfalls and snow melting periods.

The landslide’s velocity and displacement were computed for the period August 2015 to May 2023. Figure 4 shows that the average velocity in the upper (Point A) and central (Point B) parts of the slope is the highest, with an approximate velocity of 0.01 metres per day. The main direction of the landslide’s displacement is towards the North-East, consistent with the general orientation of the slope. Furthermore, a time series of the displacement indicates an apparent acceleration in the landslide’s movement since 2020, resulting in a total displacement of around 7 metres and 2.5 metres towards the North and the East, respectively (Figure 5).

A more detailed assessment was subsequently conducted in the upper zone of the landslide, where the average velocity of displacement is notably higher. As displayed in Figure 6, the magnitude of displacement in this area exceeds that of the overall landslide, with a total displacement of 10 metres to the East and 15 metres to the North. The acceleration of the displacement from 2020 is also evident at Point A.

Figure 4. Average landslide velocity displayed in the DTA platform showcasing La Barmasse slope, Switzerland. Points A and B represent the upper and central sectors of the slope, areas with particularly higher mean landslide velocities. Positive values indicate downslope velocities, and negative values, upslope velocities.

Figure 5. Time series of the landslide displacement orientation computed over the period 2015 - 2023. Showcased location: La Barmasse slope, Switzerland.

Figure 6. Time series of the landslide displacement orientation computed at Point A (upper zone of the slope) over the period 2015 - 2023. Showcased location: La Barmasse slope, Switzerland.

Case 2 - Monitoring landslide velocity and displacement in the La Valette landslide

Figure 7. The large and deep-seated landslide of La Valette in the French Alps, with a length of 2 km and an approximate volume of 3 to 4 million cubic metres. Source: Travelletti et al., 2013⁷.

La Valette landslide, which was triggered in March 1982, stands as one of the largest, and features one of the more complex slope movements in the South French Alps. This landslide reaches a length of over 2 km, with maximum depths of 25 m in the lower and central zones, and 35 m in the upper zone of the landslide. With its estimated volume of 3.5 million cubic metres, represents a significant threat to the 170 community housings located downhill in the municipality of St-Pons, Alpes-de-Haute-Provence Department.

La Valette landslide has a downhill average velocity movement spanning from 0.001 to 0.01 metres per day. The upper zone of the landslide (Point A), features the highest surface velocities. The surface velocity progressively decreases downslope to very low velocities (< 0.001 m per day) as depicted in Figure 8.

Figure 9 shows that the predominant displacement direction is towards the South-West, which is consistent with the orientation of the slope. Meanwhile, Figure 10 provides a time series from the zone characterised by higher velocity rates (i.e., Point A). In general, the dynamics of displacement are remarkably consistent between the entire slope (Figure 9) and at Point A (Figure 10). However, the magnitude of the displacement is higher in the upper zone of the landslide. The La Valette landslide recorded an approximate total displacement of up to 5 m in the South-West direction. Yet, at Point A, the total displacement reached around 25 m over the period August 2015 - March 2023.

The displacement time series of Figures 9 and 10 also suggest that the landslide displacement fluctuates from one year to another, and predominantly in the North-South component.

Figure 8. Average landslide velocity displayed in the DTA platform showcasing La Valette slope, France. Points A and B represent the upper and central sectors of the slope, areas with particularly higher mean landslide velocities. Positive values indicate downslope velocities, and negative values, upslope velocities.

Figure 9. Time series of the landslide displacement orientation computed over the period 2015 - 2023. Showcased location: La Valette slope, France.

Figure 10. Time series of the landslide displacement orientation computed at Point A (upper zone of the slope) over the period 2015 - 2023. Showcased location: La Valette slope, Switzerland.

---

⁷ Travelletti, J., Malet, J.-P., Samyn, K., Grandjean, G., Jaboyedoff, M. 2013. Control of landslide retrogression by discontinuities: evidences by the integration of airborne and ground-based geophysical information. Landslides, 10: 37-54. doi:10.1007/s10346-011-0310-8

How to visualise and analyse deep-seated landslide datasets

1. Navigate to digitaltwinalps.com and click on the ACCESS THE DTA PLATFORM button, to sign up for free. Access the platform using the link provided to you via email.
2. Proceed to select a region from the list provided on the interface. The platform offers deep-seated landslides products for two regions: La Barmasse in Switzerland and La Valette in the French Alps. For this tutorial, select La Valette.

3. After selecting La Valette region, a map interface is presented, displaying blue marker pins representing locations of ground measurement stations. Click on the pin to visualise the in-situ data captured by the stations. Alternatively, the in-situ data can also be visualised by clicking on the Deep-Seated Landslides In Situ Data icon at the bottom of the right panel.

4. Upon clicking either the blue pin or the icon a time series chart is displayed with in-situ data. Ground measurements are provided for groundwater discharge (m³/s), and daily surface displacement (m) collected at the EV2 station. Whilst at the BV16 station, you can access daily surface displacement (m), daily effective precipitation (mm), daily precipitation (mm), and daily average air temperature (^oC). Select the data of interest and adjust the time range accordingly and click Reload to visualise the measurements.

5. To visualise the deep-seated landslides products, navigate to Discover -> Disaster Risk Management -> Deep-Seated Landslides. Two product categories are provided: (a) GDM-OPT Displacement Fields (March - May 2023), and (b) GDM-OPT Time Series (2015-2023). The first category includes maps of average displacement and velocities of the landslide, whereas the second category provides displacement maps for every point in time where there is data available. Select the GDM-OPT Displacement Fields (March - May 2023) category. Proceed to adjust the time range accordingly and click Search.

6. You will notice that only one tile is available. As indicated in the previous step, this category provides average displacement and velocities for the entire period (2015 - 2023). Select the tile and proceed to Visualize. A list of the available layers is presented under the Visualize tab. Each layer is furnished with a short description and a legend. A visualisation of the product is also displayed on the map interface. Proceed to explore each of the layers, referring to the description and the legend for interpretation. For example, in La Valette landslide, a predominant displacement to the West is observed in the East-West component. The displacement is notably higher in the upper part of the landslide.

7. Gain access to time series of deep-seated landslides data by clicking back to Discover, and selecting the GDM-OPT Time Series (2015 - 2023). Once the time range is adjusted, click Search.

8. A list of tiles is presented sorted from the most recent to the oldest one. Select the latest date and click on Visualize. Navigate through the displacement maps as well as quality assessment maps (i.e., RMS maps) by clicking on the different layers and changing the dates in the Date option. The dates, highlighted with a rectangular background are the ones with available data.

9. Since this is a time series product, it enables the visualisation of a time series graph. Navigate to the Mark point of interest icon and place the yellow pin at a specific point from which time series and statistics want to be obtained. Then, proceed to click on the Statistical info icon.

10. The interface displays by default a time series of the last month. In this example, the time range is adjusted to one year. You can also select multiple layers simultaneously. Finally, export the values by using the Export CSV option.

11. In case you want access to the time series averaged over the entire landslide, click on the Create an area of interest icon and then click on Draw rectangular area of interest . Draw a polygon that covers the whole region. Next, click on the Statistical info The time series graph displayed works in the same way as it was presented in the previous step.

12. After exporting the values as CSV, the data can be further analysed. Multi-comparison at a pixel level or among different landslides can be carried out using excel or any programming language. In the image below, a Python program was used to plot the landslide displacement of the La Valette in the East-West and North-South components.

13. Proceed to explore other functionalities provided by the platform that are useful in exploring the products such as Compare different layers, Measure distance and areas, and download images. To learn more about the DTA demonstrator platform and its capabilities, follow the video tutorial accessible on digitaltwinalps.com

Try it yourself following this link