Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

This research on the sustainability of the irrigation systems is located in Flores Rancho, a community in the High Valley of Cochabamba in Bolivia. The department of Cochabamba, where Flores Rancho is located, plays an important role in the agricultural production of Bolivia. The agricultural sector employs over one fourth of the country’s population and contributed 14% to the national GDP in 2018. Over the last decades, the observed decrease in groundwater level in the study area has been raising concerns for water shortage threatening the agricultural production. It also proves that the current water use is unsustainable. Despite the initiative of the local government to stimulate more efficient pressurized irrigation systems in order to reduce water consumption, the adoption of these systems has been limited. This research quantified the impact of climate change on the local climate of the valley and identified the impacts on the local agriculture. Additionally, this research assessed the role surface irrigation practices could play in a more sustainable use of water in the area. Meteorological time series were analysed for 4 meteorological stations from 1970 until 2017 to identify any trends. This analysis revealed a large local variability of meteorological conditions and observed trends in the studied period in the valley. In all but one station, a significant warming was observed. The observed trends are higher than the expected increases in both Bolivia and the Andes, which in turn are more than double the expected global temperature increases. For all but one station, a significant drying was observed, either through an increase in reference evapotranspiration or a decrease in precipitation. These trends will increase the strain on the groundwater. For the evaluation of the irrigation and cultivation practices, the AquaCrop model, developed by FAO, was used. This model provides accurate and robust simulation of yield and water requirements for crops. As inputs for this model, interviews with farmers and field observations for soil and canopy cover were done. Actual water use of half of the interviewed farmers was observed to be higher than the annual recharge of the aquifer and therefore not sustainable. The average current water use was at least 80% higher than the water use simulated by Aquacrop. This highlights the great potential for water use reduction through improving the uniformity and timing of irrigation and the choice of planting dates. With these improvements, sustainable water use is possible with surface irrigation on all types of soils in the study area. Therefore, municipal efforts to improve sustainable water use should not only focus on high-cost sprinkler irrigation. Improved practices for surface irrigation and planting dates can also lead to short-term reduction in water use and provide additional benefits as they are low cost and relatively easy to implement.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Under the global challenge of biodiversity loss, efficient monitoring strategies should be applied to counteract the impact and enhance proper biodiversity conservation strategies. Forest biodiversity has also been affected, risking ecosystem functioning and the delivery of vital ecosystem services that they offer. This master aims to assess the feasibility of an innovative modelling approach to predict alpha diversity patterns and analyze if this alternative is suitable for monitoring and evaluating biodiversity at large scales. The major goals of this study are: Firstly, to analyze the tree species richness patterns in Belgium using spatial and statistical analysis. Secondly, to use remotely sensed data embedded in boosted regression trees to predict species richness. Thirdly, to determine the role of the variables contributing to the prediction of species richness. Lastly, to test the biodiversity-productivity and biodiversity-forest structure diversity hypotheses. The spatial and statistical analysis allowed us to identify that in Belgium, current patterns of tree species diversity are greatly influenced by different regional forest management strategies: higher diversity was found in Flanders, which has a lower proportion of coverage than in Wallonia. This was also represented in a negative relationship between increasing elevation and decreasing species richness. The study also highlights the relevance of remote sensing as a contributor to the enhanced monitoring of biodiversity. Different combinations of remote sensing data and environmental variables were used. In total, we tested 13 models, varying in seasonal remote sensing data from 2017 and 2018 and including environmental variables or not. This allowed to increase the model performance and provided a good approximation of Belgium's tree species richness spatial patterns, even though the overall accuracy was not high. Multiple factors influencing the model performance were identified, such as: sample size and the number of variables, the inclusion of environmental variables in the models, and using multiple seasons. We found that larger sample sizes provide greater performance, but also that there is the need to include multiple-seasons data, which reduces the number of samples and increases the number of predictors. We concluded that this trade-off is beneficial as phenological variations are captured according to the seasons. The inclusion of environmental variables had a positive effect on the model performance. In addition, the top predictors of diversity and variables with higher relative importance in the prediction were mainly derived from remote sensing bands instead of environmental variables. Highlighting the relevance of remote sensing data in the estimation of alpha diversity at large scales. The biodiversity-productivity and biodiversity-forest structure hypotheses are two likely explanations of how the model could relate the Sentinel-2 bands to species richness. The most relevant bands are sensitive to detect variables related to diversity such as: Leaf Area Index (LAI), Maximum chlorophyll absorption, and forest aboveground biomass (AGB). However, it is also likely that regional relationships define a large portion of the species richness outcomes.