Hieronder een korte samenvatting. Conclusies toevoegen vond ik nog lastig, maar als je dat wel graag hebt kan ik er wel nog een keer naar kijken. Verder; zijn bronvermeldingen nodig?

The Caribbean region is highly exposed to natural hazards, particularly tropical cyclones (TCs). Their impacts vary between islands, depending on hazard intensity and duration as well as local exposure and vulnerability.
A study is done focusing on the Leeward Islands (Martinique to Puerto Rico), investigating the spatial variability of three TC-related hazards: high wind speeds, storm surge, and extreme precipitation. Maximum wind speeds and their return periods are quantified using the numerical Holland (2008) wind field model. Storm surge is estimated with a simplified approximation based on the SLOSH model, which is translated into flooded areas and corresponding return periods. Total precipitation during a storm is modelled using the parametric Tropical Cyclone Rainfall (TCR) model, and return periods are determined for this hazard as well.
As input data, historical TC tracks from the IBTrACS database (1940–2024) are used, as well as synthetic tracks from STORM. All hazard modelling is carried out in CLIMADA, an open-source Python framework for climate risk assessment developed by ETH Zurich. Results are compared between islands and against regional averages, supporting the PARATUS project’s feasibility study on a regional Early Action Protocol for TCs in Antigua and Barbuda, Dominica, and Saint Kitts and Nevis.
We see that for tropical cyclone-related wind speeds, the Leeward Islands are exposed to a similar severity. For precipitation, a larger spread is found, mostly depending on the presence of mountainous regions on an island.

Cash and Voucher Assistance (CVA), a type of humanitarian aid consisting of giving money instead of products, is being used more frequently because of its effectiveness and efficiency in helping people in need (Cash Learning Partnership, 2020b). The debate on using CVA is currently focusing on improving the quality by better incorporating ’voices’ (needs and preferences) of recipients and by enhancing targeting. In targeting it is a major challenge to quickly identify the individuals and families with the biggest needs, given the lack of data (Aiken et al., 2021). Research on ways of measuring impact on and satisfaction of recipients combined with research on demographic and behavioural characteristics of recipients could lead to deeper insights in recipients of trackable CVA modalities (evouchers and ecash). This research uses the marketing literature on customer segmentation combined with machine learning algorithms to come up with an innovative new approach of categorizing recipients of evouchers, using the case of a Red Cross project on Sint Maarten. The main research question is: How can recipients of cash and voucher assistance be categorized using the field of consumer segmentation by using machine learning methods? The objective of this research is to come up with new methods to better understand recipients of CVA. Theories on customer segmentation pointed to the use of data-driven clustering methods to categorize consumers. Combined with a framework of recency, frequency and monetary aspects, recipients of evouchers could be categorized effectively. A required addition to this clustering method is to use a dimension reduction technique to avoid the negative consequence of the curse of dimensionality. Therefore, a two-step approach of dimension reduction and clustering has been applied in this research. It has been found in this research that a factor-cluster approach can lead to insightful clusters using geo-demographic data and behaviour data. Factor analysis has been used to reduce the dimensions while the k-prototype algorithm has been used to cluster into five distinct groups of recipients. The geo-demographic variables that were the most determining in characterizing distinct clusters consisted of: the age of the main beneficiary, the different household compositions and of a constructed factor ’big families and big receivers’. The most distinguishable variables on behaviour were: the number of supermarket visits (frequency), the time between the first voucher was received and the first transaction (recency) and the variables on the amount of money that was spent with the vouchers (monetary). To be able include the ’voice’ of recipients (needs and preferences), a connection between the registra tion data, behavioural data and survey data is needed. In this research only an exploratory connection could be established, due to the lack of a common identifier between the survey data and the other datasets. However, one crucial finding of this research is that it seems like the combination of these data sources can give meaningful insights in the needs, preferences and behaviour of households of Sint Maarten. With these insights specific clusters can be targeted for additional assistance, based on their needs. Recommendations for future studies include studying the validity of the found cluster results with different validation indices on cohesion and compactness, and by using simulations to determine the cluster stability. Before this factor-cluster approach can be deployed in CVA projects, more research on the treatment of limitations of this approach needs to be conducted. This is critical in communi cating the conditions and constraints of this model to humanitarian aid workers in the field. Another recommendation is to improve the design of surveys to measure the needs of recipients. For insightful factor-cluster results on the needs of recipients, survey data should be linked to geo-demographic and behaviour data. More research on including clusters in retargeting methods using feedback loops have a large potential in minimizing targeting errors and more effectively meeting the needs of recipients. With this research, the humanitarian sector can benefit from new ways to understand the needs of the most vulnerable in need. Decision-makers should build upon the feedback of recipients and move towards a new era of humanitarian assistance.

Call Detail Record (CDR) data enables the analysis of human behaviour on a large scale and the information that it contains can be promising. Not only does it allow us to track the movements of many individuals throughout time, it uncovers patterns in a persons decision making process that potentially tell us a lot about the effects of different interventions. The opportunity of finding new information on human behaviour has been noticed in several research fields, but every researcher eventually finds the same blockade: privacy. The data represents a detailed track of individuals and therefore these individuals must give approval (almost certainly lowering the amount of data that can be collected), or the data must be aggregated to the point that user anonymity is guaranteed. As a consequence of aggregated data, potentially important information could be lost. Especially in the case that both the dimension of location and time are aggregated, as these two could be considered as the essence of the CDR data. There are however techniques that increase the aggregation level, by de-aggregating the data. Naive Bayes classification has shown to be a functioning method within Machine Learning to de-aggregate a dataset that has incorporated information on at least one of the two essential dimensions; location in this case. By using the same variables to describe the administrative areas within the country that were used to describe the rows within the data, Naive Bayes classification can find the area that is most likely to fit the row. Matching the variables of the areas to the variables within the displacement dataset represents the backbone of the process, as the de-aggregation is driven by the closeness of datapoints between the two datasets.

Humans have always populated in the vicinity of river systems, where thesupply of water, nourishment and transportation is obtained from the river.However, inundation is a re-occurring problem and impact of floods are ex-pected to increase due to climate change. Accurate flood forecasting andearly warning is critical for disaster risk management. Tackling the problemof forecasting, in data scarce environments, has become increasingly impor-tant due to the changing climate. Remotely sensed river monitoring can bean effective, systematic and time-efficient technique to monitor and forecastextreme floods. Conventional flood forecasting systems require extensivedata inputs and software to model floods. Moreover, most models rely ondischarge data, which is not always available and is less accurate in a over-bank flow situations. There is a need for an alternative method which de-tects riverine inundation, using open-source data and software. This thesisaims to research the use of passive microwave radiometry for the detection,classification and forecasting of inundation.Brightness temperatures are extracted from the passive microwave radiom-etry and are converted in a discharge estimator: the C/M-ratio. Surfacewater has a low emission, thus let the C/M-ratio increase as the surfacewater percentage in the pixel increases. Sharp increases are observed forover-bank flow conditions. The research combines the identification of in-undation with a probability analysis via a quantile regressional fit. Floodforecasts can be obtained from an upstream catchment area. In the mostideal situation with a delay of2,5hours. This allows for probabilistic earlywarning decision making, with a lead time up to14days. (location specific)Strong Spearmans correlation coefficients between the discharge and C/M-ratio are found (>0.883). Allowing the model to forecast floods as gaugeddischarge records do. The model used has a comparable skill to the localGloFas forecast. This research investigated the impact the remote sensedtechnology could have on the flood forecast, response and warning system.An added model to an Early Action Protocol has the ability to lower uncer-tainty within decision making and enlarges the intervention window. Theadvice is to use such a model in combination with other forecasting modelssuch as GloFas.The challenge using this technology is the integration of hydrological com-plexity. The method allows for automated, global-covered creation of gridbased flood forecasts, independent to cloud coverage. Creating low spatialresolution flood forecasts combined with a probability bound in hours aftersatellite detection. The method has a high potential for data scarce flood-prone river basins around the world. The future for this technology lies inthe global daily availability of the data. With satellite sensors improving,spatial resolution is expected to increase. Allowing for even better floodforecasting ability.

The frequency and severity of natural disasters is increasing worldwide, leading to a growing number of people struggling to survive. While climate related natural disasters affect large portions of the world, communities who are already struggling to survive due to conflict, insecurity or poverty are hit the most. In fragile states, slowly unfolding natural disasters are getting more and more intertwined with conflict. In these areas, humanitarian and peacekeeping organizations have increasingly overlapping goals and scarce resources. Sharing information between humanitarian and peacekeeping organizations can improve the effectiveness and efficiency of both humanitarian response operations and peacekeeping missions, which may result in not only saving time and money but most importantly saving lives and reducing human suffering. Nevertheless, the process of information sharing between humanitarian and peacekeeping organizations is not common practice. This is a comprehensive study on the complexities of information sharing between humanitarian and peacekeeping organizations in fragile areas. It includes desk research, interviewing, modeling approaches and a qualitative case study on Mopti, Mali where the Red Cross Movement is actively fighting food insecurity and Dutch peacekeepers are contributing to the UN peacekeeping mission called MINUSMA.

In 2008 the Red Cross Red Crescent (RCRC) started with Forecast-based Financing pilots to improve existing Early-Warning Early Action systems. Forecast-based financing is a new methodology to prepare, deliver and respond in a more effective and efficient manner, based on hazard forecasts. Actions are triggered when a forecast exceeds a danger level in a vulnerable intervention area. Forecast-based financing consists of several implementation steps, of which the first three aim at impact-based forecasting. Therefore, In this study we investigate how forecast skill of agricultural drought forecasts can be achieved. More specifically, the aim is to identify the contribution of machine learning and satellite-derived products in early warning early action systems improving the forecast skill of agricultural drought forecasts. We explore this through a machine learning model for a case-study area of the Lower Shire River Basin in Malawi. Several experiments with different sets of predictors and predictands are conducted to test which data adds to the skill and at what spatial detail. As predictors, the following agro-climatic indices are used: cumulative precipitation, soil moisture anomalies,mland surface temperature anomalies, El Niño Southern Oscillation in July and four different dry spell categories within the growing season (0-2 days dry spell, 3-4 day dry spell, 5-10 day dry spell and larger than 10 day dry spell). As drought predictand, the normalized difference vegetation index (NDVI) and the vegetation optical depth
(VOD) in March are used, the latter obtained from satellite data company VanderSat. The final set of predictors and predictands is narrowed down based on which data is available and with which quality (timeliness, reliability, accuracy). Initial results, show higher accuracy and weighted accuracy values for the models including soil moisture data compared to the ones without soil moisture, expect for the last month in the growing season, where it give opposite results. The outcome of the model can support humanitarian organisations to increase the lead time necessary to act upon a drought trigger and reduce the impact of such event.

In an emergency, humanitarian organisations share information to prevent redundant data collection and avoid gaps and overlap in the relief activities that they undertake. An analysis of hygiene kit distribution in the Bangladesh-Myanmar displacement crisis and consultation of both literature and humanitarian professionals led to the construction of a model on information diffusion in complex emergencies. This model proved to be able to evaluate strategies that have a level of complexity that could not be apprehended by existing models. Experimentation with this model leads to the conclusion that a locally sourced team, with an outward focused organisation that produces near real-time information products, is most effective in diffusing information.