Choose an application

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

Tropical forests are facing threats due to climate change; therefore, it is crucial that we can understand how these forests responded to past climate events to be able to predict the future effects of climate change. Dendrochronology, the study of tree rings, has been one effective way that past climate effects have been reconstructed, with many recent advances happening in the field of tropical dendrochronology. Unfortunately, dendrochronology is difficult to perform on a large scale, due to the time and resources needed to perform sampling of trees. Satellite imagery could be a way to track tropical tree growth variation on a larger scale. Links have been found between tree ring measurements and satellite-derived metrics in temperate regions already. The goal of this research was to compare satellite-derived metrics and ring width indices (RWI) in tropical forest regions on a local and pan-tropical scale. The local analysis was performed on a tree ring data set from a region near the Paru river in Brazil on a per-pixel basis. We used a linear mixed effects (LME) model to compare RWI with three satellite-derived vegetation indices at different time shifts. These were the yearly maximum with no shift, three month shift, six month shift and one year shift, and the yearly summed integral with no shift for the enhanced vegetation index (EVI), the normalized difference vegetation index (NDVI) and the gross primary production (GPP). The analysis on a global scale compared tree ring datasets from five different tropical regions, this time taking an average of the satellite derived metrics over the study areas to compare with RWI. The local Paru analysis showed significant effects of MODIS-derived GPP max and GPP sum with a one year time shift, and of Landsat-derived EVI and NDVI sum with no shift in determining RWI. For the pan-tropical analysis no satellite-derived metrics showed to have a significant effect on RWI. This shows that it is possible to compare satellite imagery and RWI on a local scale, but not across separate tropical regions.

Choose an application

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

Arbuscular mycorrhizal fungi (AMF) enhance the resilience and sustainability of banana (Musa spp.) cropping systems by increasing nutrient and water uptake. Moreover, these fungi interconnect banana plants and link them with intercrops through a common mycorrhizal network (CMN). Successfully establishing these networks in banana plantations requires additional insights in mycotrophic intercrops that initiate the CMN as well as in alternative field inoculation methods. We researched the effect of the AMF species Rhizophagus irregularis on plant growth of the banana cultivar ‘Williams’ and on intercrops by means of pot and field experiments in Valverde Mao, the Dominican Republic. Growth variables and biomass production were determined and mycorrhizal root colonization quantified. In the field inoculation experiments, changes in the carbon and nitrogen content of the banana leaves and in culturable soil and root-associated microorganisms were investigated. Moreover, the interactions between R. irregularis and the plant growth promoting organisms Bacillus amyloliquefaciens and Trichoderma harzianum were researched on banana plantlets. This was followed by PCR detection of Bacillus spp. in the rhizospheric soil and by PCR identification of local AMF species from root samples. Our research suggests that velvet bean (Mucuna pruriens) is the most promising initiator of CMNs in banana plantations. After inoculation, a positive trend in this legume’s growth was detected. In addition, a non-inoculated central banana plantlet acquired a significantly higher root colonization through the velvet bean’s CMN, than through the CMN of an inoculated banana plant. However, reasons for the absence of banana growth promotion in this treatment should be examined more thoroughly. The experiment on the single or co-inoculation of AMF, B. amyloliquefaciens and T. harzianum did not yield conclusive results on biomass development. However, mycorrhizal root colonization was not reduced in the co-inoculated treatments. PCR detection indicated the local presence of Bacillus spp. and R. irregularis strains. The soil injection of R. irregularis did not show significant effects on banana tree growth, leaf nutritional status, microorganisms’ composition nor root colonization. This is more likely due to the short experimental period of eight weeks than due to failed inoculation. As the use of velvet bean to create a CMN is promising and questions remain on the possibilities of AMF soil injection, long-term experiments in banana plantations are proposed.

Choose an application

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

Tropical highland ecosystems have been relatively understudied in terms of their carbon balance, despite their potential importance as carbon stores and the multiple other ecosystem services they provide. In this thesis, we explored soil carbon stocks and sources, as well as greenhouse gas exchange between soils at the atmosphere, in the Aberdare range (Kenya), at altitudes between 2009 and 3461 m. The average soil organic carbon (SOC) stock in the upper 30 cm across all sites was 153.1 ± 42.9 t ha-1. With these SOC stocks, the soils in this area contained a significant percentage of Kenya's total SOC storage (between 1.34 and 1.44 %), despite its limited surface area (~ 0.3 %). SOC stocks were found to be mainly dependent on altitude, with an increase of 18.4 t ha-1 per 1000 m. Furthermore, our results along with earlier data collected in the same project, showed that more carbon is stored in the top 30 cm of the soil in the alpine and subalpine vegetation and the high-altitude shrub zone than in the montane forest and b bamboo zones. The local vegetation supplies the majority of carbon to the SOC stocks. Across the Aberdares range, the vegetation appeared to be follow predominantly the C3 photosynthetic pathway. The lower temperatures and sufficient precipitation favored plants with this photosynthetic pathway over plants with a C4 pathway. However, in profiles on the central plateau of the range, between 3000 and 3300 m, there is clear evidence of a vegetation shift from C4 to C3 vegetation in the past, with many soil profiles showing strong shifts in stable carbon isotope ratios (δ13C) in deeper layers, with an increase of δ13C values of 6 ‰ or more. This shift was mainly present in profiles under Erica vegetation and alpine and sub-alpine vegetation, but also some profiles taken in Hagenia forests showed a shift. From the literature it could be deduced that this shift probably occurred about 5000 years ago. In addition, a shift from C4 to C3 vegetation was also found in the dry evergreen forests in the north of the park. The cause here might be a land use change from pasture to forest. Soils acted as a net sink of atmospheric methane (CH4) in the the vast majority of sites (79 out of 93 flux measurements). The uptake rates of CH4 varied between -0.20 and -6.83 nmol m-2 s- 1. The dominant process behind this uptake is the oxidation of CH4 by methanotrophic bacteria in the aerobic zone of the soil. On the other hand, a net emission of CH4 from the soil was observed in 14 of the 93 sample collection points. This occurred mainly at the wet subsites of 4 the sites in peatland vegetation, and resulted from methanogenesis in the anoxic layer of these highly organic soils. Moreover, the detailed porewater profiles of CH4 concentrations and their δ13C values taken at these sites indicate that more CH4 is produced at these sites than is actually emitted by the soil towards the atmosphere. Stable isotope data on CH4 and CO2 demonstrated this to be caused by methane oxidation by methanotrophic bacteria in the shallower layers of the soil. In terms of CO2, net emissions from the soil were observed at all sites, as expected, and varied much less strongly than CH4 fluxes. In general, lower fluxes were observed at sites in the alpine and sub-alpine vegetation zone and the montane forest zone. In addition, there was a significant negative correlation (R²=0.381 at a significance level of 0.05) between the magnitude of CO2 flux and altitude. Other pa

Choose an application

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

Semi-natural grasslands in Europe are subject to many sources of pressure. Degradation, agricultural intensification, land use change and climate change all have a very strong impact on these systems. The ecosystem services associated to semi-natural grasslands make them important features in the fragmented landscape of Western Europe. Adequate conservation strategies are thus required for their survival. The Natura 2000 network of the EU has created an internationally acknowledged legal base for protection of these habitats, among others, but the monitoring methods that are currently applied to these systems are often inefficient. The recent launches of the Sentinel-2 satellites of the Copernicus program may pose a viable alternative to the traditional monitoring efforts. The applicability of multispectral remote sensing to observe the environment has been thoroughly explored with other systems, such as the Landsat fleet. The technological improvements of Sentinel-2 could make it a useful tool for continuous monitoring of natural habitats. However, its applicability for assessing the conservation status of temperate semi-natural grasslands in Western Europe has not yet been adequately investigated. In this thesis, freely available Sentinel-2 imagery of spectral reflectance in three periods in the growing season is used to develop models for regression analysis of five traits that are related to conservation status of semi-natural grasslands. These traits are aboveground dry biomass, leaf area index, vegetation water content, canopy nitrogen content and canopy phosphorus content. The regression analysis was performed with linear models, partial least squares regression and decision trees and kernel-based machine learning algorithms. The Random Forests method gave the best results for all traits (overall adjusted R2 > 0.60 for all traits, > 0.70 for structural traits, nRMSE < 0.41 for all traits, 0.22 for LAI), but differences between areas and periods were significant. It seems to be difficult to determine an overall best period for estimation of the traits for each area separately. However, models with datasets of all areas combined also gave adequate correlations with relatively low errors. All bands seemed to be important for the models of each trait, reflecting the relatively low amount of information in single bands. It would therefore be interesting to enrich the models with more information, such as vegetation indices. Monitoring of semi-natural grasslands with Sentinel-2 is thus indeed an alternative that is worthy of further investigation. Improvements to these models are however necessary before application in real conservation schemes. Temporally more precise data and larger datasets, especially for the biochemical traits, can probably attribute a lot to the accuracy of the models.

Choose an application

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

The impact of climate change on the environment has caused significant challenges and opportunities for various industries, including the wine industry. The Belgian wine industry has shown immense growth in recent years, but the optimization and discovery of viticultural techniques adapted to the Belgian climate are still required. This thesis focused on enhancing berry yeast assimilable nitrogen (YAN) concentrations by implementing three different cultivation techniques. Lowering the leaf wall height of Souvignier Gris grapevines by 1/3 (from 150 cm to 100 cm) resulted in increased trends of ammonia (65%), alpha-amino nitrogen (18.2%) and total berry YAN concentrations (18.8%), due to the dilution of the total obtained nitrogen among a smaller plant volume. Moreover, berries consistently contained less soluble solids and experienced a delay in veraison-onset, attributable to the decline in photosynthetically active canopy. By implementing a second technique, it was established for Chardonnay grapevines, that berry YAN concentrations rise in response to the foliar application of Ferticell (3.2 kg N/ha) and urea (15 kg N/ha), with a combination of both showing the highest efficiency (+145%) due to their synergistic relationship. Furthermore, fermentation rates were positively affected by an increased dose of nitrogen application. A final approach on Cabernet Cortis vines included applying different nutrition at various development stages through fertigation. It was revealed that vines provided with the most amount of NH4+, at the expense of the more preferred NO3-, produced shoots with shorter internodes and heavier bunches. Both can be related to other vines experiencing situations of excessive nitrogen. Additionally, vines receiving the most nitrate were characterized by the highest must YAN concentration (397.67 mg/L). Sensory evaluation of Chardonnay and Cabernet Cortis wines revealed that wine created from the must with the highest initial YAN was perceived as the most fruity and flowery, and additionally seemed to be the most preferred. This is attributable to the positive effect of amino acids on the formation of sensory metabolites associated with these aromas. In conclusion, it can be stated that under Belgian climate conditions, lowering the leaf wall height by 1/3, applying a combination of urea and Ferticell (18.2 kg N/ha) to the foliage and providing vines with nitrogen nutrition through fertigation, are all sufficient approaches to increase the YAN in berries at harvest. Due to a very dry summer, total YAN concentrations remained generally low and the optimal level of YAN could not be identified.