Senegal

The first step was to find research papers that could guide us to start this project efficiently considering the fact that we only had two months to implement a solution. The most interesting papers we found and used are: County-Level Soybean Yield Prediction Using Deep CNN-LSTM Model, Deep Gaussian Process for Crop Yield Prediction Based on Remote Sensing Data, and Deep Transfer Learning for Crop Yield Prediction with Remote Sensing Data.

Data

So the first question answered thanks to these papers was: What data do we need to train a Deep Learning model predicting crop yield? 

Indeed, the authors used two types of raw data: 

• Remote sensing data downloaded with Google Earth Engine (GEE)
• Ground truth crop yield data: we had yield data collected by IPAR for the production of maize, rice, and millet in 2014

So, we downloaded the datasets MOD09A1.006 Terra Surface Reflectance 8-Day Global 500m and MYD11A2.006 Aqua Land Surface Temperature and Emissivity 8-Day Global 1km for the regions and departments of Senegal using Shapefiles. The first dataset has 7 bands of surface spectral reflectance that can be used to calculate the Normalized Difference Vegetation Index (NDVI), an indicator of vegetation’s health. The NDVI is calculated from the red light (which vegetation absorbs) and near-infrared light (which vegetation strongly reflects) reflected by vegetation. 

The second dataset has two bands: temperature during the day and temperature during the night. 

Which Crop Land Cover should we use? What is the Crop Land Cover used for?

The Crop Land Cover dataset is used as a crop cover mask. This means that all pixels of the reflectance and temperature images that are not classified as cropland pixels will be removed from the images so that the model will only be exposed to data from the crops, and not from cities for example. 

In the end, we used a different Land Cover dataset than the papers above. After comparing the MCD12Q1.006 MODIS Land Cover Type Yearly Global 500m that these papers used with the Copernicus Global Land Cover Layers, we decided to use the latter. This decision was taken after comparing the datasets with cropland maps of Senegal:

Here, we see that the crops are mostly located in the South-West and the Northern regions of Senegal. However, when we look at the MCD12Q1.006 MODIS dataset where the cropland is labeled as brown, we see that a lot of the cropland is actually missing (Northern crops and most of the rice crops next to the Casamance River) but this is not the case with the Copernicus dataset, where the cropland is in labeled as pink. So we came to the conclusion that the Copernicus dataset was the most accurate for Senegal.

The only problem with the Copernicus dataset is the time range: 2015–2020, knowing that we had ground truth yield data for the year 2014. However, we assumed that the land cover for the year 2014 was close enough to the one from 2015 and still better than the 2014 MCD12Q1.006 MODIS dataset, so used the 2015 Copernicus land cover as a crop mask for images from 2014.

At the end of this step, we had collected data for the three GEE datasets presented above, for the entire country, the regions, departments, and GPS locations from the IPAR study.  

Preprocessing of the Data

According to the papers cited previously, using 3-D pixels count histograms instead of raw satellite images for the prediction of yield helps to avoid the model from overfitting (model too closely fit to a limited set of data points).

What is the number of weeks mentioned above? 

In order to focus on the growing season, we only studied the images of the weeks following the planting of the seeds to the harvesting, which is, for instance, week 19–30 for maize. 

Deep Learning Model 

We decided to use the Deep Learning architecture from this paper which is a CNN-LSTM: 

“CNN can learn the relevant features from an image at different levels similar to a human brain. An LSTM has the capability of bridging long time lags between inputs over arbitrary time intervals. The use of LSTM improves the efficiency of depicting temporal patterns at various frequencies, which is a desirable feature in the analysis of crop growing cycles with different lengths.”

We also tried to use the CNN architecture proposed in this paper but the results were not satisfying, the model was finding a random value for all data points that minimized the loss, and then did not learn or improve afterward even with different hyperparameters. Since we were having better results with the CNN-LSTM, we decided to only focus on the latter. 

Transfer Learning 

We used transfer learning to improve the maize model. We had some yield data from South Sudan and Ethiopia (Source: deep-transfer-learning-crop-prediction) that we used to train the model and then fine-tuned it using the yield data from Senegal. 

Data augmentation

We tried to do some data augmentation on the IPAR dataset by taking sliding windows around the point of origin (lat/lon) and assuming the yield of the crops in these sliding windows was the same as at the point. 

This method did not improve the maize and millet results but did improve the rice model.

Results 

We ran several trainings with the different approaches explained previously and collected the resulting metrics:

We can see that the Transfer Learning for the maize model did improve the MSE (Mean square error) and therefore was our best MSE. In comparison, the millet model did not do as well as the maize model but we did not have any other data to perform transfer learning. Finally, the rice model could be improved using data from other countries from the same Github repository where you found the maize data. To be noted: the MSE is higher for the rice model because the yields of rice are higher than the maize and millet yields in the first place (up to 14 T/ha). 

Here are some visualizations of the maize results: 

We also ran the predictions on every department of Senegal over 4 years (2015–2018) for maize, rice and millet: 

Final product

We created an interactive notebook where the user can select the region they want to predict the yield. The user can also choose the year and crop type. While this notebook has several application areas. For instance, it can be used as a tool for policymakers to decide what food to import and export in order to maintain food security in the country. The tool can also help farmers make management and financial decisions.

We also implemented another notebook that will take as input the GPS latitude and longitude instead of a selected area:

Conclusion 

To conclude, in these two months, we were able to implement a Deep Learning model that predicts crop yield in Senegal following this schema: 

As mentioned in this article, the lack of ground truth data was an issue that made the models not as efficient and accurate for Senegal as they could be. An improvement easily implementable would be to have ground-truth data like the 2014 IPAR dataset but for several years in order to be able to show the model the fluctuations over the years so that it could learn it and be more flexible to variations in the data.

The Niayes region in northwestern Senegal forms a coastal strip which has particular climatic conditions that allow intensive cropping and cattle breeding even during the dry season. In addition to local breeds, important non-native cattle populations are maintained for milk production. The area is also known for high densities of horses which are of high economic value and used to transport food crops. However, the cost effectiveness of maintaining these animals has been continuously threatened by exposure to African animal trypanosomiasis (nagana), a wasting disease transmitted by the tsetse fly and considered one of the most important constraints to cattle production in infested areas. In 2006, a parasitological and serological survey of resident cattle showed up to 90% herd prevalence rates of the disease. 

Eradication campaigns are essential tools to control such diseases. However, they are expensive and if the target fly populations are not truly isolated from each other, cleared zones are gradually reinfested by tsetse flies from neighboring areas and the eradication campaign fails. The first attempt to eliminate tsetse flies from approximately 150km of habitat in the region was undertaken in the 1970s using selective bush clearing and residual ground spraying. While no flies were detected immediately following the campaign, they reappeared in the 1980s. A second campaign was then launched using insecticide spraying and traps. However, the flies re-emerged by the late 1990s with severe consequences for livestock and farmers’ livelihoods. A key reason for the failure of these efforts is that the campaigns were initiated when only half of the infested areas had been identified and the eradication activities did not reach the entire target population. 

Data as the missing link

Given the isolated nature of the Niayes tsetse population, resurgence after the two campaigns in the ‘70s and ‘80s described above were due to population build-up from small residual pockets inside the Niayes area. This indicates that both efforts lacked comprehensive and precise data about the spatial distribution of the flies within the region, limiting the ability of the activities to target the entire population.

Following learnings from the previous campaigns, the Government of Senegal initiated a new tsetse eradication project in the Niayes region in 2005. This project started with a feasibility study in 2006 to assess the possibility of creating a tsetse-free zone in the Niayes region. The study was supported by the International Atomic Energy Agency (IAEA), the Food and Agriculture Organisation of the United Nations (FAO), the International Cooperation Centre of Agricultural Research for Development, and the Government of Senegal through the Senegalese Institute for Agricultural Research and the Directorate for Veterinary Services. Following a period of trials, training, and preparation, beginning in 2012, the project implemented the sterile insect technique, which involves releasing mass-produced sterile male flies in the infested areas to halt reproduction and ultimately eradicate the flies. This method is considered one of the most environmentally friendly tactics. The results of this project fit into a FAO/IAEA pan-African tsetse fly eradication campaign launched in 2001.

The success of this project was driven by the precision and accuracy of the feasibility study conducted from 2006-2010. Specifically, the study used innovative data and methods including satellite data, geographic information systems, and mathematical models to identify the spatial distribution of the flies and confirm isolation of the target population. These data and methods allowed the project to increase efficiency by covering a large area with few resources, reduced costs, and improved granularity and accuracy of insights, ensuring that the eradication strategy was targeted and thorough. The study also confirmed that the tsetse population in the Niayes region was isolated from the southern part of Senegal by dry areas where the fly could not survive. This meant that, once the flies were eliminated in the identified areas, the risk of re-infestation was very low, ensuring success of eradication activities.

“As these [tsetse flies] are very sensitive to environmental conditions, the collection of temperature, rainfall, vegetation, and other environmental data was essential. Satellite data offers the solution to provide information on this type of parameters.” – Dr. Assane Fall, Institut Sénégalais de Recherches Agricoles, Department of Bio-Ecology and Parasitic Pathologies 

Tsetse flies cannot survive without suitable habitats, including forest tree cover and presence of fresh water. This means that identifying environmental conditions in a given area would indicate the presence of flies as well as the potential for development of flies. Satellite data allowed the team to identify vegetation and wet areas across the seasonal cycle through a singular data source. In the absence of satellite data, the team would have had to use a combination of ground surveying, potentially using drones to map the area, and sensors to collect data on temperature, humidity, rainfall, vegetation, and other characteristics. Even with these multiple methods, it would have been difficult to cover the entire area, particularly at the level of precision required. “So, using satellite data has the advantage of collecting all the data we need through a single download,” explained Dr. Fall.

The wet area classification from the satellite data enabled the project to detect tsetse flies in unexpected sites, such as military camps where sampling would not have been implemented otherwise. In addition, some areas identified as wet areas were remote and would not have been covered through traditional ground surveys. During implementation, these areas required the use of GPS to locate the sites that were sometimes several kilometers away from established tracks. 

“Satellite data has the advantage of offering a wide range of data types. In addition, they made it possible to have an acceptable precision for our project in order to prevent the omission of favorable areas capable of harboring flies that can recolonize the area,” says Dr. Fall. 

The target zone for the project covered 1000 km2. Using satellite data and modeling techniques, the study was able to restrict the area that needed to be surveyed to 4% of the total surface area covered. This meant significant cost savings for data collection activities. Use of satellite data and these innovative methods were estimated to have reduced the sampling costs by 90%, while optimizing the efficiency and accuracy of the insights.

Sampling strategy used in the Niayes of Senegal to delimit the target population of G. p. gambiensis (source).

Impact

In an FAO report on the project, Baba Sall, then Project Manager and Head of the Animal and Health Section in the Ministry of Livestock at the time, said, “Eradicating the flies will significantly improve food security, and contribute to socio-economic progress.” 

A cost-benefit analysis of the project found that the eradication of the tsetse fly in the Niayes region would result in annual benefits of 2.8 million Euros. The cost of the project was estimated to be 6,400 Euros per km2. However, the study found that this was significantly outweighed by the additional income of 2,800 Euros per km2  to farmers on an annual basis, meaning that the project would pay for itself in 2.5 years. Research estimated that income of the rural population would increase by 30%, as farmers would be able to sell more milk and meat.

The eradication project has allowed farmers to switch from less productive disease-resistant cows to more productive imported breeds. Before the successful eradication of the flies, farmers kept local breeds despite their low milk and meat production and low reproductive rates because they are naturally tolerant to trypanosomiasis. However, now they are replacing breeds that produce 1-2 liters of milk a day with imported breeds that can produce between 20-40 liters. In addition to reducing the cost of treating animals, farmers have substantially increased both productivity and income. The improved milk and meat production of those breeds has enabled farmers to maintain smaller herds, reducing the grazing pressure on the fragile ecosystem without sacrificing income. 

Loulou Mendy, a pig farmer in the area quoted in an FAO publication said, “Life has become more comfortable not only for the animals, but also for the farmers, now, we can even sleep out in the open. This was unthinkable before because of the tsetse bites.”

Within Senegal, Dr. Fall and his team report that the benefits of satellite data have been recognized and have since been used in multiple other projects to model animal diseases. In particular, it has been used to map the risk of diseases such as Rift Valley fever, foot-and-mouth disease, and the avian influenza to help decision-makers design and implement interventions.  Internationally, the geospatial data has been incorporated into similar projects in a number of countries including Uganda.

Dr. Seydina Ousmane Sene contributed research for this data story.

Between October and December 2020, experts from Togo, Guinea, and Senegal met via Zoom for capacity-strengthening activities organized by the Global Partnership for Sustainable Development Data (Global Partnership) and led by the United Nations Environment Program and the Danish Hydraulic Institute (UNEP-DHI) on the UNEP-DHI Flood and Drought Portal, an Earth observation- (EO-)based tool. The Flood and Drought Portal is a series of open-source technical applications that help stakeholders to perform baseline assessments using readily available satellite data, impact assessments through data analysis, planning options, and dissemination of environmental data to affected groups or individuals, allowing authorities to be better prepared to act early, anticipate the risks of extreme weather events, and empower citizens to avoid flood prone areas.

Senegal joined the training to increase its capacity to use satellite imagery to produce reliable data for environmental protection and climate change adaptation policies. The participants came from the Directorate of Water Resources Management and Planning, the National Agency for Statistics and Demography, the Ecological Monitoring Center, the Agricultural and Rural Prospective Initiative (IPAR), and the Department of Planning and Environmental Watch (DPVE) of the Ministry of the Environment and Sustainable Development of Senegal. Following the three-session training, participants identified a use case for their newly acquired skills in their respective countries. The Senegal team, including Mbengue, decided to tackle the issue of flood management.

Cheikh Faye, statistical engineer at IPAR, noted that in Senegal, “local actors and humanitarian organizations often have to work with traditional meteorological sensors (ground instruments capable of detecting temperature, relative humidity, atmospheric, wind direction, speed and radiation, for example) and climate data that is dated, poor quality, or nonexistent, due to the limited ability to access and work with EO data in near-real time.” In addition to data, indicators are needed to allow the combination or interpretation of data as a basis for making sound decisions or policies and to help compile complex values ​​into a simple interpretation of a particular situation, for example, conditions and types and severity of drought or flooding. These indices are linked and a combined analysis of the relevant scenarios measured enables one to effectively define the flood risks for a given area.

The Senegal team assessed a range of flood indicators available on the portal, including the Standard Precipitation Index, the Normalized Antecedent Precipitation Index, Soil Health Index, Flood Variability Index, and Flood Risk Index. The team worked on these indicators and evaluated the Flood and Drought Portal’s algorithms with reference to the country's geospatial data and different eco-geographic zones. The objective of this assessment was to validate the availability and accessibility of near-real-time EO flooding data with ground-level data, with the hope of justifying the use of EO data in decision-making processes for predicting and addressing environmental risks.

Capacity-building results

Through the Flood and Drought Portal, the Senegal team discovered that several indicators can be used to improve existing knowledge on flood-prone areas and effectively identify areas at risk of flooding to inform early warning systems, as well as urban planning. From the team’s point of view, the potential of the information drawn from the Portal presents a case for the development of a spatial mapping database and a decision support tool for all stakeholders in Senegal to access the data, visualizations, and reports on flood risks in the country.

For example, viewing the Soil Moisture Index in Senegal through the Portal shows that soil moisture decreased from south to north and from west to east across the country, reflecting the national records of floods from 2007 to 2020, where the north has a longer dry season and little to no flooding with annual rainfall of up to 380mm, while flood incidences increase as they move south and east, where the rainfall can reach up to 1,500 mm per year. Compared to ground measurements, the remote sensing technique as demonstrated by the Flood and Drought Portal can capture soil moisture information across the country in a single snapshot, which takes less time and labor. The data, which also complemented traditional rainfall data over the same period, can help predict flooding based on predicted rainfall relative to soil moisture levels.

The Senegal team observed some limits with the portal due to a lack of data disaggregated at regional and departmental levels. With a global tool, it wasn’t feasible to generate small-area statistics for all administrative areas across the country in the Flood and Drought Portal, especially since Senegal wasn't a country that the UNEP-DHI team had initially worked with in developing the tool and delivering training activities. The data and indicators in the Portal are available at resolutions based on satellite sensors (e.g., the soil water index is at a 0.1-degree spatial resolution). However, the Portal’s map features offer the possibility to disaggregate all indicators at the regional level that can then guide supplementary analysis at sub-regional levels.

According to Mbengue, the training enabled the team to acquire new knowledge on climate indicators. The open-source nature of the Portal, which allows data exports to different platforms with lower-level metrics for further analysis, is also a very useful feature, further supporting the need for the country to invest in such a platform that can inform authorities as well as the public about flood risks, as this does not yet exist in the country. “We have learned to explore and use data on deforestation, drought, and flooding in Senegal, which is not often collected at the national level. New indicators were discovered in the platform with concepts and methodology thanks to the existing documentation,” he said. "This has allowed us to better understand the concepts and methods that are used to measure or appreciate the extent of deforestation, flooding, or drought and to strengthen analyses on these indicators. In addition, we were able to obtain a new source of data on these various national issues.”

Mbengue added that the team had acquired a baseline to easily assess the effectiveness of environmental projects that will be implemented in the future. This saved the team time and resources to collect this data, which is easily accessible on the portal. “The idea was to collect data to improve the foundations of knowledge,” said Mbengue. “It turned out that we could not focus on the small administrative areas of the city, but the aim was to improve the knowledge base, and we have now acquired the ability to analyze new indicators of flooding in flood-prone areas in order to anticipate and warn about the risks of flooding.”

Next steps

The DVPE AND IPAR plan to present their findings and recommendations to the Civil Aviation and Meteorology Agency, the Ministry of Water and Sanitation, and the Ministry of Regional Planning, as well as other technical and financial partners, in a bid to build partnerships for the development of the flood observatory system, which can be based on the data provided by the Flood and Drought Portal in conjunction with more granular field data and provide relevant analysis based on a range of indicators and indices.

Once the information obtained has been well refined and analyzed, the Senegal team will be able to make evidence-based recommendations to the authorities for the implementation of the country’s Ten-Year Flood Management Program (2012-2022). On the training, Faye observed: “The training allowed us to acquire new knowledge. Even though some platforms could not inform us of some more granular indicators that are relevant for the country, this was useful due to the rapid availability of this information which can easily be transferred and used in other software.”

The Global Partnership is organizing a knowledge sharing event to support the dissemination of lessons and recommendations from this training through facilitation of the three countries involved in the training and peer countries and stakeholders by the end of December 2021

This project was managed by Francois Kamano, Francophone Africa Program Manager, and the case study was written by Muthoni Mugo, MEL Program Officer.

Entre octobre et décembre 2020, des experts du Togo, de la Guinée et du Sénégal se sont réunis virtuellement via Zoom pour le renforcement des capacités dirigé par le Programme des Nations Unies pour l'environnement et l'Institut hydraulique danois (PNUE-DHI) sur un outil basé sur les observations de la Terre pour la gestion des inondations, la gestion de la sécheresse, surveillance agricole et planification de la sécurité de l'eau (le portail PNUE-DHI sur les inondations et la sécheresse), organisé par le Global Partnership for Sustainable Development Data (le Global Partnership). Le portail Inondations et sécheresse est une série d'applications techniques open source aidant les parties prenantes à effectuer des évaluations de base à l'aide de données satellitaires facilement disponibles, des évaluations d'impact par l'analyse des données, des options de planification et la diffusion de données environnementales aux groupes ou individus concernés, permettant aux autorités d’être mieux préparé pour agir tôt, anticiper les risques d'événements météorologiques extrêmes et donner aux citoyens / résidents les moyens d'éviter les zones sujettes aux inondations.

Le Sénégal a rejoint la formation pour accroître sa capacité à utiliser l’imagerie satellitaire pour produire des données fiables pour la protection de l’environnement et les politiques d’adaptation au changement climatique. Les participants étaient issus de la Direction de la gestion et de la planification des ressources en eau (DGPRE), de l'Agence nationale de la statistique et de la démographie (ANSD), du Centre de suivi écologique, l’Initiative de prospective agricole et rurale (IPAR) et de la Direction de la planification et de la veille environnementale (DPVE) de la Ministère de l'Environnement et du développement durable du Sénégal. À la suite de la formation en trois sessions, les participants des trois pays ont identifié un cas d'utilisation pour leurs compétences nouvellement acquises dans leurs pays respectifs. L'équipe du Sénégal, dont Mbengue, a décidé d'aborder la question de la gestion des inondations.

Cheikh Faye, ingénieur statisticien à l'IPAR, a noté qu'au Sénégal , « les acteurs locaux et les organisations humanitaires doivent souvent travailler avec des capteurs météorologiques traditionnels (instruments au sol capables de détecter la température, l'humidité relative, la pression atmosphérique, la direction du vent, la vitesse et radiations, par exemple) et les données climatiques qui sont datées, de mauvaise qualité ou inexistantes, en raison de la capacité limitée d'accéder et de travailler avec des données d'OT en temps quasi réel » . En plus des données, des indicateurs sont nécessaires pour permettre la combinaison ou l'interprétation des données comme base pour prendre des décisions ou des politiques judicieuses et pour aider à compiler des valeurs complexes en une interprétation simple d'une situation particulière, par exemple les conditions et les types et la gravité de sécheresse ou d'inondation. Ces indices sont liés et une analyse combinée des scénarios pertinents qu'ils mesurent permet de définir efficacement les risques d'inondation pour une zone donnée.

L'équipe du Sénégal a évalué une gamme d'indicateurs d'inondation disponibles sur le portail, y compris l'indice de précipitation standard (SPI), l'indice de précipitation d'antécédent normalisé (NAPI), l'indice d'humidité du sol (SHI), l'indice de variabilité des inondations et l'indice de risque d'inondation. L'équipe a travaillé sur ces indicateurs et évalué les algorithmes de la Plateforme en référence aux données géospatiales du pays et aux différentes zones éco-géographiques. L'objectif de cette évaluation était de valider la disponibilité et l'accessibilité des données d'inondation d'observation de la Terre (OT) en temps quasi réel avec des données au niveau du sol, dans l'espoir de justifier l'utilisation des données d'OT dans les processus politiques/de prise de décision pour prévoir et traiter les risques environnementaux.

Résultats du renforcement des capacités

A travers le Portail Inondations et Sécheresse, l'équipe du Sénégal a découvert que plusieurs indicateurs peuvent être utilisés pour améliorer les connaissances existantes sur les zones inondables et identifier efficacement les zones à risque d'inondation, afin d'informer les systèmes d'alerte précoce pour les inondations, ainsi que la planification de l'urbanisation. Du point de vue de l'équipe, le potentiel des informations tirées du portail présente un cas pour le développement d'une base de données de cartographie spatiale et d'un outil d'aide à la décision pour toutes les parties prenantes au Sénégal pour accéder aux données, visualisations et rapports sur les risques d'inondation dans le pays.

Par exemple, la visualisation de l’indice d’humidité du sol du Sénégal sur le portail montre que l'humidité du sol a diminué du sud au nord et d'ouest en est à travers le pays, reflétant les records nationaux d'inondations de 2007 à 2020 où le nord a une saison sèche plus longue et peu à aucune inondation avec des précipitations annuelles allant jusqu'à 380 mm, tandis que les incidences d'inondations augmentent en se déplaçant vers le sud et l'est, où les précipitations peuvent atteindre jusqu'à 1 500 mm par an. Comparée aux mesures au sol, la technique de télédétection telle que démontrée par le portail Inondations et sécheresse peut capturer les informations sur l'humidité du sol à travers le pays en un seul instantané, ce qui prend moins de temps et de travail. Les données, qui étaient également concurrentes aux données pluviométriques traditionnelles sur la même période, peuvent aider à prévoir les inondations sur la base des précipitations prévues par rapport aux niveaux d'humidité du sol.

L'équipe du Sénégal a observé certaines limites avec le portail en raison d’un manque de données désagrégées au niveau régional et départemental. Avec un outil mondial, il n'était pas possible de générer des statistiques sur les petites régions pour toutes les régions administratives du pays dans le Portail sur les inondations et la sécheresse, d’autant plus que le Sénégal n'était pas un pays que l'équipe PNUE-DHI avait d’abord travaillé avec elle à l'élaboration de l'outil et à la prestation des activités de formation.  Les données et les indicateurs du portail sont disponibles à des solutions basées sur des capteurs satellites (par exemple, l’indice d’eau du sol est à une résolution spatiale de 0,1 degré). Cependant, les caractéristiques cartographiques du portail offrent la possibilité de désagréger tous les indicateurs au niveau régional qui peuvent ensuite guider l’analyse supplémentaire aux niveaux sous-régionaux. 

Selon M. Mbengue, la formation a permis à l'équipe d'acquérir de nouvelles connaissances sur les indicateurs climatiques. La nature open-source du portail qui permet l'exportation des données vers différentes plateformes avec des mesures de moindre niveau pour une analyse plus approfondie est également une fonctionnalité très utile, soutenir davantage la nécessité pour le pays d'investir dans une plate-forme spécifique au pays qui peut informer les autorités ainsi que le public sur les risques d'inondation, car cela n'existe pas encore dans le pays. « Nous avons appris à explorer et à utiliser les données sur la déforestation, la sécheresse et les inondations au Sénégal, qui ne sont pas souvent collectées au niveau national. De nouveaux indicateurs ont été découverts dans la plateforme avec des concepts et une méthodologie grâce à la documentation existante », a-t-il déclaré . “ Cela nous a permis de mieux comprendre les concepts et méthodes qui sont utilisés pour mesurer ou apprécier l'ampleur de la déforestation, des inondations, ou l'intensité de la sécheresse et de renforcer les analyses sur ces indicateurs. De plus, nous avons pu obtenir une nouvelle source de données sur ces différents enjeux nationaux.

Mbengue a également déclaré que l'équipe avait acquis une base de référence pour évaluer facilement l'efficacité des projets environnementaux qui seront mis en œuvre à l'avenir. Cela a permis à l'équipe d'économiser du temps et des ressources pour recueillir ces données, qui sont facilement accessibles sur le portail. « L'idée était de recueillir des données pour améliorer les fondements de la connaissance », explique M. Mbengue. « Il s'est avéré que nous ne pouvions pas nous concentrer sur les petites zones administratives de la ville, mais l'objectif était d'améliorer la base de connaissances, et nous avons maintenant acquis la capacité d'analyser de nouveaux indicateurs d'inondation sur les zones inondables afin d’anticiper et alerter sur les risques d'inondation”.

Prochaines étapes

La DVPE et l'IPAR prévoient présenter leurs conclusions et recommandations à l'Agence de l'Aviation Civile et de la Météorologie (ANACIM), le Ministère de l'Eau et de l'Assainissement, le Ministère de l'Aménagement du Territoire ainsi qu’a d'autres partenaires techniques et financiers, afin de construire des partenariat pour le développement du système d'observatoire des crues qui peut s'appuyer sur les données fournies par le portail Inondations et sécheresse en conjonction avec des données de terrain plus granulaires et fournir des analyses pertinentes basées sur une gamme de indicateurs et indices.

Une fois que les informations obtenues auront été bien affinées et analysées, l'équipe du Sénégal pourra formuler des recommandations fondées sur des preuves aux autorités pour la mise en œuvre du Programme décennal de gestion des inondations du pays (2012-2022). Sur l'ensemble de la formation, M. Faye a observé : « La formation nous a permis d'acquérir de nouvelles connaissances. Malgré le fait que certaines plateformes ne pouvaient pas nous informer de certains indicateurs pertinents pour le pays, cela a été vraiment utile en raison de la disponibilité rapide de ces informations qui peuvent facilement être transférées et utilisées dans d'autres logiciels.

Le Global Partnership organise un événement de partage des connaissances pour soutenir la diffusion des enseignements et des recommandations de cette formation grâce à la facilitation des trois pays impliqués dans la formation et des pays pairs/parties prenantes vers la fin décembre 2021.

Francois Kamano, Francophone Africa Program Manager, a géré ce projet et Muthoni Mugo, MEL Program Officer, a écrit cette étude de cas.