Featured Thesis Figure

The Impact of the Urban Heat Island Effect on Mortality in the City of Bern

During my master’s thesis, I investigated the influence of the urban heat island effect on mortality in the city of Bern. The urban heat island effect describes the phenomenon of increased temperatures in urban areas in comparison to their rural surroundings. This in turn has a strong impact on people's well-being, as the human body needs to cool down at night to recover from the heat stress of the previous day. We therefore linked high-resolution temperature and mortality data based on the study subjects' places of residence. In a next step, we used a case-crossover study design and conditional logistic regression and distributed lag non-linear models to analyse the association between increased temperatures and mortality risk. The results of such analyses are visualized with exposure-response functions (ERFs). These show the temperature (exposure) on the x-axis and the corresponding relative mortality risk (RR, response) on the y-axis.

This key figure shows the ERFs for nighttime temperature (A), for daytime temperature (B) and for urban heat island intensity (C). Starting from the minimum mortality temperature (MMT, set at 18 °C for Bern), a temperature at which the mortality risk is minimal, we were able to determine an increase in RR with increasing temperature for all three temperature variables. For better comparability, the 99^th percentile temperature (extreme temperature, which occurs in 1 out of 100 cases) is shown as a vertical dashed line. The grey areas represent the confidence intervals. The graph is to be interpreted in such a way that the mortality risk in (A) increases by 45% after a night with an average temperature of 23.3 °C (99^th percentile temperature) compared to a night with a temperature of 18 °C after which the risk is minimal. For a daytime temperature of 25.8 °C (B) the mortality risk increases by 31% compared to the risk at 18 °C and for a UHII of 2.4 °C (C) the risk is 25% higher than at a UHII of 2 °C. If we compare the risk increase of the nighttime temperature with that of the daytime temperature, it becomes apparent that the risk increases more at elevated nighttime temperatures than at daytime temperatures. On the one hand, this may indicate a lack of nocturnal recovery from the preceding heat stress for the human body; on the other hand, this additional increase could result from the urban heat island effect. The combination of urban climate and health data is highly relevant to prepare our cities for a future in an even warmer climate and to recognize which specific risk groups need special attention.

Figure from the master thesis by Adrienne Wehrli, 2024.

Read more about student research

Investigating the planetary boundary layer height at the CLOUDLAB field site

The planetary boundary layer (PBL) is the bottom-most layer of the atmosphere. It is directly influenced by the Earth’s surface and responds to surface forcing within about an hour. The flow in the PBL is characterised by turbulence during the day, which means that the PBL is usually well-mixed. The PBL is also the part of the atmosphere in which humans live. Determining its properties and particularly its height (PBLH) is therefore of great relevance. For example, the PBLH can influence air quality by controlling the volume of air into which pollutants released near the ground are mixed. In my project, I investigated the PBLH at the CLOUDLAB field site in Eriswil, Switzerland. The goal was to compare instrument and method performances under different atmospheric conditions.

The colour shade in the figure shows the attenuated backscatter measured by a ceilometer. Ceilometers are usually used for measuring cloud base heights (white triangles) but the attenuated backscatter can also be used as a proxy for aerosol concentration. The aerosol concentration is expected to be higher in the PBL than in the free troposphere above, so the strong change in backscatter can be used to estimate the PBLH. A number of algorithms exist to detect this change: red circles show the PBLH determined by the manufacturer algorithm installed on the ceilometer, whereas black and yellow circles show PBLH estimations from the open-source research algorithm ‘STRATfinder’.
The PBLH can also be estimated from temperature and humidity profiles. Orange symbols show a detected temperature inversion, where the temperature increases with height. Temperature inversions form a cap on vertical mixing so can be used as a marker of the PBLH. Blue symbols show the PBLH estimated from the relative humidity profile. The relative humidity is expected to be higher in the well-mixed PBL than in the free troposphere, so a minimum in the vertical relative humidity gradient can be used as a marker of the PBLH. The PBLH can also be detected from a maximum peak in the potential temperature gradient (purple symbols). This method is based on the expectation that the free troposphere is more stably stratified than the PBL.

I applied these thermodynamic methods to temperature and humidity profiles measured by radiosondes (diamond symbols), unmanned aerial vehicles (triangle symbols) and a microwave radiometer (square symbols). A microwave radiometer is a remote sensing instrument which measures microwave emission from molecules in the atmosphere and uses this to infer the temperature and humidity profiles.
I compared the different methods and instruments using case study days such as the one shown, as well as systematic comparisons over a year of observations. On this day, the thermodynamic and aerosol methods show good agreement. There is a well-defined PBLH with a clear temperature inversion and strong aerosol concentration gradient. However, on a systematic level, the agreement between the PBLH estimated by different methods and instruments was rather poor. In particular, the presence of low-level clouds can lead to non-ideal PBL structure and evolution, in which case the features detected by the thermodynamic and aerosol methods are often not at the same height.

Figure from the master thesis by Heather Corden, 2023.

Read more about student research

Linking Seasonal Dynamics of Snow-Covered Area to Wet-Snow Avalanche Activity

This thesis aimed to improve snow-cover models by validating model output (Jules Investigation Model JIM_OSHD) with webcam data (fractional snow-covered area; fSCA) and snow-water equivalents (SWE) from the Dorfberg site in Davos, Switzerland. For the winter seasons 2015 / 2016 to 2019 / 2020 we established a link between fSCA data, SWE and wet-snow avalanche activity, and developed a simple model to detect the onset of wet-snow avalanche activity cycles.
We find that the webcam fSCA technique is most reliable for detecting such cycles at high elevation zones during peak winter. In contrast, the JIM_OSHD model overestimated fSCA as there was a substantial offset between JIM_OSHD fSCA and webcam fSCA. We showed that webcam fSCA-data could be a reasonable technique to validate snow cover models or could even be an alternative to snow cover models when analysing seasonal trends where site specific and high spatial resolution is needed.
By linking seasonally changing fSCA to the onset of wet-snow avalanche cycles we confirmed our hypothesis that wet-snow avalanches cycles are likely to be released when both fSCA and SWE have reached their maximal value prior to the ablation stage in spring. In addition, our simulations of the historic onset of wet-snow avalanche cycles worked well for extensive and winters with average or above average snow depth and less well for short winters with snow depths below average.

Figure from the thesis "Detecting Avalanches: Linking Seasonal Dynamics of Snow-Covered Area to Wet-Snow Avalanche Activity" by Flavia Mäder, 2020.

Read more about student research

Response of lake stratification to forest density

The response of lake stratification to forest density was assessed for Lago di Mezzano in Central Italy and the past 18’000 years. We used laminated lake sediments as a “historical diary” to assess past environmental conditions. Pigments present in purple sulphur bacteria and stored in the sediment served as an indicator for anoxic conditions up to the photic zone. In addition, preserved tree and shrub pollen were used as a proxy of forest density. We found that closed forests correlate with low oxygen availability (anoxic conditions) and lake stratification, while open landscapes correlate with lake mixing, see the Figure. In a forested landscape (upper panel), the lake surface is shielded from wind, which reduces lake mixing. This means that the lake becomes stratified more easily. Open forests at Lago di Mezzano are related to human activity since the early Neolithic. Such open landscapes (lower panel) favour lake mixing (black arrows lower panel) and a cease of anoxia because of higher wind shear. These oxic conditions promoted sedimentary phosphorous trapping (red arrow lower panel) and reduced the lake productivity. On the other hand, closed forest stands increased anoxia, phosphorous recycling (red arrows upper panel), and lake productivity (e.g. benthic and planktonic diatoms). Thus, a return to a forested natural state today could lead to lake stratification at Lago di Mezzano. Similar relationships between forest density and lake mixing have been observed in several small and deep temperate and Mediterranean lakes during the Holocene. We therefore suggest that landscape management plays an important role in how lake ecosystems respond to ongoing climate change.

Figure from the thesis "High-resolution record of primary productivity and anoxia in the context of the environmental history of Lago di Mezzano, Central Italy, since the Late Glacial" by Luc Hächler, 2021.

Read more about student research

Noble Gases as Mean Ocean Temperature Proxies in an Earth System Model

The oceans are the major heat reservoir of the Earth system and play a critical role in regulating its surface temperatures. In face of the on-going anthropogenic impacts on the climate system, understanding the feedbacks in ocean-atmosphere heat exchange is of particular importance. To this end proxies have been developed that enable reconstruction of the global mean ocean temperature (MOT) and thus provide quantitative estimates of the past global radiative balance.
Noble gases were proposed as MOT proxies due to their chemical inertness and temperature-dependent solubilities: they are more soluble in colder water, which hence reduces the atmospheric inventory, and vice versa. Therefore, noble gas mixing ratios of the past atmosphere preserved in ice cores can be used to reconstruct past MOT. In this project, the implementation of the noble gas krypton (Kr) in the Bern3D Earth system model was tested under different climatic and oceanic boundary conditions. The goal of these comprehensive sensitivity tests was to assess the reliability of noble gases as MOT proxies.
The upper two panels illustrate the Kr saturation ratios of simulated oceans under different wind stress. In these two different scenarios with the same MOT, different atmospheric Kr inventories were obtained due to different Kr saturation ratios in the oceans. The dark blue region in the top panel indicates large volumes of undersaturated water in the deep ocean, while the oceans in the middle panel are around 99% saturated with Kr almost everywhere. This difference is caused by sea ice blocking the surface air-sea gas exchange but not the heat exchange, which leads to a decoupling of the noble gas tracer from the temperature signal. This means under the situation of large area of sea ice existing, especially over the Antarctic bottom water formation region, calibration is required for MOT reconstruction using noble gas proxies.
In the Bern3D model control simulation, sea ice is set to have zero leakage and completely prohibit any surface gas exchange, which does not describe the reality precisely. To better understand the impacts of ice leak, sensitivity tests were run with ice leak values from 0 to 1, corresponding to an air-sea gas exchange range from 0 to 100%. The bottom panel shows that the deficit caused by sea ice on the atmospheric Kr inventory decreases non-linearly with the increase of ice leak. The difference in δKratm is 0.22‰ between ice leaks of 0.2 and 1, close to the current ice core noble gas mixing ratio measurement uncertainty (0.2‰). Currently a no-sea-ice situation is assumed when interpreting the ice core measurements. While noble gases remain reliable as MOT proxies, our quantification on sea ice impact shows the necessity of taking calibrations in future measurements.

Figure from the thesis "Noble Gases as Mean Ocean Temperature Proxies in Preindustrial and Last Glacial Maximum Simulations" by Zhijun Liu, 2021.

Read more about student research

Studying the causes of Recurrent Rossby Wave Packets

The causes of Recurrent Rossby Wave Packets (RRWPs), which can lead to long-lasting and possibly high-impact surface weather were investigated for the period of 1979 to 2018. Composite and regression analyses revealed atmospheric blocking, changes in the background zonal wind, and forcing from deep convection as potential drivers of RRWPs. The relationships between the identified potential drivers and RRWPs were quantified and visualised with so-called causal networks using the causal inference method PCMCI. Using iterative conditional independence tests, PCMCI reveals the causal structure between time series.
The figure shows a causal network derived with PCMCI for RRWPs in the North Atlantic. Each orange to red circle represents one variable at its approximate location. The colour of the circle denotes its autocorrelation (labelled as autocorr. path coefficient in the legend). The red and blue lines represent the statistical links between the variables (path coefficient). Directed arrows indicate a causal link from one variable to another and lines ending in white circles denote correlations without directionality. The colour of the arrows shows the sign and strength of the links, red arrows standing for a positive link and blue arrows for a negative one. The black numbers on the links are the time lag at which these links were found. This model used 3-day timesteps, therefore a time lag of “1” corresponds to a link within a time lag of 3-5 days (lag “0” is days 0-2).
The causal network shows that RRWPs, denoted by R, are increased by blocking in the North Atlantic (Blocks N Atl) and decreased by blocking over Scandinavia (Blocks Scand), both at time lag 2. The blocking, however, is driven strongly by a decrease in the local zonal background wind at lag 1 and instantaneously (see the blue links between U250 N Atl and U250 N Eu to the blocking variables). And finally, RRWPs (R) drive a decrease in the local background wind at different lags. A similar pattern was also found for RRWPs in the eastern North Pacific and using different temporal resolutions. The evidence for forcing from deep convection acting as a driver of RRWPs is less clear and the links are susceptible to change depending on temporal resolution of the model. However, it is shown that RRWPs can drive the atmospheric circulation elsewhere and increase deep convection, as the link from R to OLR Gulf Guinea shows; lower values of Outgoing Longwave Radiation being a proxy for deep convection.

Figure from the thesis "Analysing Potential Drivers of Recurrent Rossby Wave Packets using Causal Inference" by Lukas Meyer, 2022.

Read more about student research

Simulations of heat waves in Bern in 2018 and 2019 with an urban climate model

Three heat waves in Bern in 2018 and 2019 were simulated with the numerical urban climate model MUKLIMO_3. The model output was validated against low-cost measurements of air temperature from more than 70 locations in the greater Bern area.
The figure shows the distribution of mean air temperature three meters above ground at 02:00 am during the first heat wave in 2019 (city of Bern in the center; buildings in black). Shown is the temperature difference between the location of the official MeteoSwiss weather station in Bern/Zollikofen (black cross) and each grid cell of the model. The urban city centre is modelled up to 1.6 K warmer than the location of the rural weather station. The formation of an urban heat island is thus evident in the model. Forested areas and areas at higher elevations show lower air temperatures. The Aare River shows no cooling effect on air temperature due to very high water temperatures (dark red grid cells surrounding the city center).
Our model underestimated the highest observed air temperatures during the day, and overestimated the lowest observed air temperatures during the night. In particular, the model showed slower cooling in the first half of the night compared to the measurements.
Despite such limitations, MUKLIMO_3 is able to reproduce the basic distribution of air temperature in the greater Bern are during heat waves. The model performs particularly well for heat waves on calm days without precipitation. Hence, it can provide initial indications of heat stress with relatively little effort, and can be a basis for urban planning measures..

Figure  from the thesis "Assessment of the urban climate model MUKLIMO 3 for three heat waves in Bern" by André Hürzeler, 2021.

Read more about student research

Coalition formation among industrialised and developing countries with and without the implementation of a Border Tax Adjustment

Can Border Tax Adjustments (BTA) facilitate and improve the formation of International Environmental Agreements (IEAs), in the sense that more countries accept more stringent emission reductions? To answer this question, two policy scenarios were derived from a game-theoretic model; in the first one, countries introduce a domestic producer tax on the production of the so-called dirty good responsible for GHG emissions, and in the second one, a classic BTA is implemented, where this domestic tax is extended to traded dirty goods.
The figure shows the coalition formation among industrialised and developing countries with and without the implementation of a BTA. The stability functions have to be positive for both types of countries such that they want to be part of the coalition and have no incentive to leave it. The x-axis indicates the number of industrialised countries (in grey) in the coalition, hence coalition size minus this number indicates the number of developing countries (in blue) in the coalition. For up to three countries in the coalition, a stable coalition can be formed under both policy scenarios. A higher coalition size can only be reached under the BTA scenario. The largest achievable coalition consists of six members, of which one is an industrialised country and five are developing countries. Coalition sizes above six continue to be profitable for industrialised countries with a BTA in place, but because of the welfare losses of the developing countries they cannot sustain a larger stable coalition unless they would "pay" the developing countries for their membership. We find that with such a transfer scheme in place, a coalition size of eight members, consisting of three industrialised countries and five developing countries, is stable. Our results point out an important factor for international climate negotiations: the inclusion of developing countries in climate negotiations (through the UNFCCC principle of "common but differentiated responsibilities") because it opens up the opportunity for substantial efficiency gains.
The thesis has further shown that a BTA exhibits some advantages over current climate policy instruments, since it is able to tackle key concerns raised by countries planning to move forward with unilateral climate policies such as carbon leakage and competitive losses. Furthermore, the implementation of a BTA facilitates the formation of IEAs under which especially industrialised countries accept more stringent emission reductions than with a simple domestic producer tax, such that global emissions can be reduced. Therefore, the concept serves as a complementary policy tool for national and regional climate strategies with the aim of helping to reach climate goals formulated under the Paris Agreement.

Figure  from the thesis "International Environmental Agreements and Trade Sanctions: A Game-Theoretical Analysis of the Impact of a Border Tax Adjustment" by Sophie Kaufmann, 2021.

Read more about student research

Numerical reconstruction of daily weather from the Societas Meteorologica Palatina network.

This thesis sought to recover and evaluate the instrumental weather data in the yearly publications of the Societas Meteorologica Palatina. Data rescue efforts revealed the overall good quality of the observations. In a second part, we reconstructed daily temperature and sea level pressure (SLP) fields for Europe for the extreme winter 1788/89 on a 0.1x0.1 horizontal grid. The reconstructions were computed using the historical station data from the Palatina network and an analogue resampling method (ARM), then further improved with an ensemble Kalman fitting (EnKF) procedure. The figure depicts daily fields of SLP at the end of December 1788. The panels for 24-25 December show strong SLP gradients and a pronounced cyclonic system over north-eastern Europe, which led to a spell of stormy weather and brief warming. Blocking over Scandinavia around 27-29 December implied a northeasterly Arctic air flow making its way into Central Europe and a return to subzero temperatures.
Results from validation tests showed the very good predictive skill of the reconstructions, which is promising for further numerical analyses of daily weather patterns at a regional scale, and notably for periods in the 18th century.

Figure from the thesis "Unlocking weather data from the Societas Meteorologica Palatina (1781-92): a daily reconstruction of the winter 1788/89" by Duncan Pappert, 2021.

Read more about student research

Predicted yield density distribution of winter wheat in Switzerland under increased temperatures

This thesis used correlation analysis to study which climate indicators relate most strongly to winter wheat yields in Switzerland from 1981 to 2017, and constructed a multiple regression model to predict yield on the basis of climate variables. The model’s most important climate variables were temperature, especially summer heat and winter frost. Adverse effects of warmer winter temperatures and precipitation were also identified. The model was also used to predict a yield density distribution based on data 2 °C and 4 °C warmer. This distribution showed that the mean and median yield might decrease only very slightly at higher temperatures. However, the spread increased at higher temperatures, so minimum yields were lower and maximum yields were higher. This increased spread is likely due to the model including both positive and negative effects of temperature, so the resulting yield might vary more widely from year to year. This may indicate that in a warmer future, farmers may need to adapt in order to avoid years with large losses.

Figure 8 from the thesis “Attribution of winter wheat yield variability to climate drivers in Switzerland” by Raeleigh Price, 2020.

Read more about student research

Paleo-climatic influence on flat-topped carbonate platforms in Lac de Joux, Switzerland

Based on previous seismic and bathymetric analyses of the Lac de Joux, Switzerland, this study addresses the nature of isolated flat-topped carbonate platforms in 1 to 5 m water depths that are not attached to the shore.
By reviewing the bathymetry (panel A) and the seismic stratigraphy (panel B), we interpreted the platforms as structures on nuclei formed by past glaciations. Indeed, glaciers in the Joux Valley shaped retreating moraines into drumlins (1). These geomorphic analyses were followed by sediment analyses that investigated the sediments covering these glacial features: due to the change of temperatures from the Pleistocene to the Holocene, carbonate precipitated and accumulated onto the platforms (3). Visible on the seismic sections, windward progradation (3b) of the platform succeeded accumulation (3a) as the sediment supply exceeded the accommodation space below wave base. Our analysis shows that these carbonates were chemically precipitated and bio-induced by algae, mainly charophytes. Additionally, periods or events of clastic spills into the region of the platforms were confirmed by the higher Ca/K ratio and Fe content. Together with the changes in sedimentary calcite concentration, this allows us to state that paleo-climatic changes in the environment surrounding the platforms influenced them directly.

Figures 18 and 19 from the master thesis by Eléonore Perret.

Read more about student research

Vertical particle export during the last glacial cycle in Southern Indian Ocean sediment cores

The oceanic sediment core MD12-3394 was retrieved in the Southern Indian Ocean. This region is characterized by a very high biological productivity in the surface ocean. Accordingly, the particle accumulation rate on the sea floor is very high. For this Thesis, the cycles of Marine Isotope Stage 3 (MIS 3) were analyzed. These cycles were rapid climate fluctuations between 35 and 60 kyr BP. The Figure shows: (A) the barite flux in the MD12-3394 core, which is a proxy for export production; (B) the atmospheric CO₂ concentration reconstruction from the EPICA Dome C ice core (Bereiter et al., 2015), correlated with climate variations; and (C) the lithogenic flux, representing the amount of terrestrial material brought to the core site.
During past ice ages, large atmospheric CO₂ concentration variations were closely linked with biological productivity and eolian dust input in the Southern Ocean. It is apparent from this figure that whenever the biological productivity was increased, the atmospheric CO₂ concentration increased simultaneously, and the lithogenic flux decreased. This implies that the biological productivity at this location was not a control on atmospheric CO₂ concentration; as we would have expected an increased primary production to consume CO₂ and transfer it to the depth, therefore reducing its concentration. Additionally, the decrease in lithogenic flux, concomitant with the increased biological productivity, indicates that eolian dust fertilization is not a factor.
As a result, the only process capable of increasing biological productivity at the ocean surface in this situation is an increased upwelling. Increased upwelling would strengthen the biological pump, but reduce its efficiency. The net outcome is a stronger biological productivity, and an outgassing of deeply sequestered CO₂. This points to atmospheric CO₂ regulation mechanisms during the small scale variations within the glacial cycle, similar to those identified for large glacial to interglacial CO₂ variations. Master thesis by Antoine Thévenaz.

Read more about student research

Homogenised monthly series from the meteorological diaries of Johann Rudolf von Salis-Marschlins 1781-1800

The meteorological diaries of Johann Rudolf von Salis-Marschlins contain a rich collection of climatological data. Within the studied years of 1781-1800, well over 10’000 pressure and temperature measurements, more than 4’000 observations of phenological phases, and close to 2’000 descriptions of precipitation events can be found. So far, the Marschlinian diaries have played a minor role in historical climatology, which is mainly due to reported inaccuracies of the used instruments. With the construction and homogenization of a monthly temperature series (upper panel), pressure series, and precipitation depth totals, this master thesis sheds new light with robust, quantitative information on local climate during this early-instrumental period. Further potential of the Marschlinian diaries is evaluated by means of an analysis of snow cover data and observations of phenological phases (see 12 selected phases in bottom panel, cf. Figures 3.13. and 4.4., pages 74 and 112, from the master’s thesis by Markus Grimmer).

Read more about student research

Swiss cantonal strategies for adaptation to climate change

Using a Qualitative Comparative Analysis, this thesis contributes to answering the question why a Swiss canton has developed a strategy in order to politically tackle climate adaptation. Such a strategy is the dependent variable (y-axis in the Figure, label ADAPT). A score of 1 means that a canton has elaborated an adaptation strategy (e.g. canton of BE). A score of 0.67 stands for cantons that plan to elaborate an adaptation strategy but have not done so yet (e.g.  canton of VD). Scores below 0.5 indicate that adaptation strategies are neither planned nor elaborated. In total, sufficient explanations (so-called solution paths, displayed on the x-axis) as to why cantonal adaptation strategies are formulated are: (1) defining climate adaptation as a legislative objective, (2) emphasizing strong mitigation policies together with the presence of a left-wing minister, and (3) emphasizing mitigation policies together with the existence of a high problem perception (shown here, label MIT*ALPS*PERC). Climate adaptation and mitigation, however, do not always occur in parallel: defining adaptation measures does not automatically imply that mitigation measures are also given high priority. Master thesis by Eva Wieser.

Read more student research

Simulating the potential range of shallow landslides

Shallow landslides pose a considerable danger to humans and infrastructure due to their sudden triggering and high velocity. In this master thesis, the CLIMADA2 model was implemented to simulate the risk for shallow landslides. The model is designed to be globally applicable with minimum data requirements and incorporates an algorithm to determine possible source areas and flow paths of shallow landslides. The flow path is assessed by an algorithm adapted from Flow-R which consists of a flow direction, inertia and friction model. Its output grid can be interpreted as a point’s, spatial relative, probability to be reached by a shallow landslide (see Figure). A comparison of modelled with observed lengths of shallow landslides revealed that the flow path algorithm is not suitable to model single shallow landslides. However, the model captures the overall distribution of slide lengths and is expected to be skillful for modelling extreme (long) slides. (Figure from the master thesis by Thomas Rölli, 2018)

Read more student research

Strategic interaction between countries can hinder incentives to invest in climate-related technologies

Climate-related technology can be an option to keep global temperature increase below the 2° C target set by the Paris Agreement.
This master thesis tests whether strategic interaction (interaction between n>2 players, who inter alia maximise their respective utilities, and where the players’ payoff depends on the actions of the other players) hinders or fosters countries (players) to invest in four climate-related technologies, (i.e., abatement AB, energy efficiency EE, adaptation AD and low-carbon technology LC) in a non-cooperative setting.
It is shown that strategic interaction hinders the incentives to invest in abatement technology, fosters the incentives to invest in energy efficiency and adaptation technology, and can both foster and hinder the incentives to invest in low-carbon technology. This is true for all countries. We conclude that this could partly explain the low level of current mitigation measures worldwide.
Specifially, the strategic interaction effect (the product of the marginal damage and the total derivative of the aggregate emissions of all the other countries with respect to a technology parameter, see Figure) hinders the incentives to invest in AB and LC technologies and fosters the incentives to invest in EE and AD technologies for the global top-5 CO2 emitters. Alos, the direct effect (the partial derivative of a benefit, cost, or damage function with respect to a technology parameter) always outweighs the indirect, i.e, the strategic interaction effect. Hence, none of the countries has an incentive to invest in a less efficient AB or LC technology due to strategic reasons. However, in the case of symmetrical countries, a country would always prefer to invest in a less efficient abatement technology due to strategic reasons.  This implies that if the countries are homogeneous enough, they might want to commit to lower mitigation levels in order to pass the burden of mitigation to the other countries.  In the context of international climate negotiations, this implies a free-riding strategy on abatement technology.. (Figure from the master thesis by Anna Kulakovskaya, 2018)

Read more student research

Glacier-friendly weather?

In this thesis, a number of historical air temperature and pressure measurements from European stations were used to answer this question with regards to the Lower Grindelwald Glacier. To assess potential north-to-south shifts of storm tracks, a 2–6-day bandpass Lanczos filter was applied to the pressure data (see Figure). It was found that a run of cold summer seasons between 1840 and 1847 and many cold and wet fall seasons between 1836 and 1853 were linked to a strengthening and a southward shift of the storm tracks. We argue that this contributed strongly to the last growing phase of the Lower Grindelwald Glacier in the mid-19th century. (Figures 17 and 19, pages 58 and 60, from the master thesis by Julian Flückiger)

Read more student research

Recurving tropical cyclones

Tropical cyclones are known for their severe direct impacts. Some of them undergo extratropical transition when they recurve (from north-westerly to easterly directions) over the Atlantic and move poleward into a midlatitude environment. They have a potential for high impact weather further downstream, over Europe.

During extratropical transition, the ascending air and enhanced moisture transport causes strong latent heat release and inject air with low potential voriticity into the upper troposphere (upper panel of the Figure). This in turn accelerates the upper level jet and amplifies its waviness.  This meridional amplification of the extratropical flow over the Atlantic is important for the formation of extreme precipitation events over Europe (lower panel of the Figure).

This thesis provides the first climatological assessment of the effect of recurving North Atlantic tropical cyclones on precipitation extremes over Europe. We find a significant increase in precipitation extremes over Europe in three out of four major configurations of the atmospheric flow, and between 30 and 72 hours after the interaction onset. The response is however sensitive to the jet configuration at the moment of the interaction.

Species richness in a warmer climate

Agricultural intensification and climate change are among the main drivers for biodiversity changes.
In this thesis, a two-part hurdle model was implemented to predict present and future plant richness in grasslands of the western Swiss Plateau. An extensification of grasslands (>=20% of pastures area from intensive to extensive) under warmer conditions (+ 2° C in mean winter temperature) is found to produce higher species richness (i.e. up to 70 % more species) in the study area until 2050.
The southern Broye valley, the pre-alpine zones and today's high-diversity areas (yellow to green shades) would profit most from less intense pasture management combined with higher winter temperatures. Conservation measures would therefore be most effective in these regions. (Figure from the master thesis by Luca Bussani)

Read more student research

Drivers of Inter-Annual Variability in Wood Density

Dendroclimatology utilizes a number of tree-ring parameters to reconstruct past climates. This thesis investigates the mechanisms that drive variations in wood density of eight Pinus cembra cores, using an image analysis software with very high resolution.
A biplot of the first two principal components (PC) shows the relationships between relevant anatomical features of tree rings (shown are abbreviations in the Figure). The first PC (x-axis; explaining 47% of the variance in the original dataset) can be seen as a measure of  the cell wall thickening phase. The second PC (y-axis; 19% explained) is a good measure of mean and maximum density. The vector length is equal to the variance of the variables. Furthermore,  the distance between vectors refers to the similarity between variables. In short,  latewood width (LWW; highlighted) is strongly correlated to density parameters (MDCWT & MDCWA), i.e., closely resembles mean density.
This suggests that latewood width can serve as a surrogate for density if such measurements are not attainable. Understanding how anatomical features of tree rings are related to density parameters helps to better understand the mechanisms that drive inter-annual variations in wood density. (Figure 34, page 55, from the master thesis by Lisbeth Haberbauer)

Read more student research

Reconstruction of historical flood levels of the river Aare in Aarau

Over the past centuries, the Swiss city of Aarau has been recurrently affected by floods. In this thesis, historical floods are assessed by means of administrative protocols, chronicles, maps, paintings, and newspaper articles. Between 1400 and 1900, a number of 39 flood events were dated (not shown). Prior to 1700, floods occurred significantly earlier in the year than after 1700, a seasonal pattern that is in accordance with observations of rivers in Germany. Water levels of ten of these events were reconstructed for the first time in this thesis (orange dots for events prior to 1852, the bars indicate probable ranges). Data after 1852 are adjusted gauge readings. Local river regulations were rather ineffective, and recurring conflicts between involved political levels delayed effective measures until the 1850s. On a regional scale, the upstream Kander deviation in 1714 and the Jura water correction 1868-1891 had a great impact on the river system in Aarau. They substantially deepened the riverbed and thus decreased the danger of floods (effect approximated by the two black dotted lines). (Figure 17, page 85, from the master thesis by Mauro Bolzern)

Read more student research

This decade's solar photovoltaic emergence: a cross-sectorial investigation

Solar photovoltaic technology has been available since the 1950’s however the technology has not been widely applied until the beginning of this decade.  In this thesis, the paramount factors for this boom are investigated using Qualitative Comparative Analysis of the leading promoting countries, i.e. with at least 10 gigawatts of installed solar photovoltaic capacity.

The analysis revealed three potential pathways. The first positive pathway (Path 1 in the Figure; dot sizes represent importance) includes the cases of Japan and the United States. They are representative of cases that have "Ability" and* "Resource Intensive Populations" and* "Potential to decarbonize their electricity and heat production sectors". The second pathway (Path 2), represented by China and Japan, is similar to the first in the sense that they have the same parsimonious solution (taking only the large circles into consideration). The third and final pathway (Path 3) is indicative of a more sustainable path taken by European countries such as, Germany and Italy. This more sustainable pathway is highlighted by the presence of "Kyoto Commitments" and* the "Inability to rely on nuclear power production in the future" and* "Potential to decarbonize" and* "Ability".

Path 3 is in essence the most desirable because it encompasses emission reduction pledges under international climate negotiations. In contrast, Path 1 and 2 seem to be primarily fueled by large, resource intensive populations and growing economical necessities. (Figure 15, page 50, from the master thesis by Simon Steffen).)

Read more student research

Changes in the risk of spring frost damage for grapevines due to climate change
A case study in the Swiss Rhone Valley

Global warming does not only cause the growing phases of plants to start earlier, it also leads to a shift in the average day of year of latest frost days. Depending on the magnitude of these changes, as well as on the corresponding variances, it is not clear if risk for spring frost damage will increase or decrease. Focusing on the Swiss Rhone Valley, future risk according to the SRES A1B scenario was modelled by combining two locations, four frost indices, twelve phenology models and 16 climate model chains. The solid lines indicate the yearly means of the frost indices (i.e. the probability of a frost day after budburst (BBCH09), p(T.n<0°C), the frequency of frost days after BBCH09, f(T.n<0°C), the temperature sum of daily minimum temperature below 0 °C after BBCH09, ∑(T.n<0°C), and the temperature sum of daily mean temperatures below 2 °C after BBCH09, ∑(T.0<2°C)), the day of year (DOY) of the last frost day and of BBCH09. The shaded areas indicate the range of +/- one standard deviation around the yearly means. The broken red lines visualize the mean trends for the period 2001 to 2050. The average risk for spring frost damage is predicted to slightly decrease. This may support winemakers in their decisions regarding future grape varieties and frost damage prevention measures.
(Figure 12, page 66, from the master thesis by Michael Meier)

Read more student research

Quinoa yields w.r.t. temperature and precipitation (Lampa, Arapa, and Chuquibambilla stations, Peru, years 1964 - 2012)

The potential value of seasonal climate forecasts for allocating agricultural land is estimated in this thesis. Therefore, a theoretical model was developed and applied to the region of Puno in Peru. The application relies on using statistical climate data and simulated yields for quinoa and potato crops - the Figure presents yield simulations for quinoa depending on two climatic indicators, i.e. the seasonal precipitation sum and the minimum temperature during the flowering stage. The analysis shows that the expected revenue can be increased by 4.66 % when key decisions are made with seasonal forecasts, compared to a scenario without seasonal forecasts. The positive value indicates the potential of integrating high-quality seasonal forecasts into decision-making when allocating agricultural land to specific crops.
(Figure 4, page 24, from the master thesis by Anita Frehner)

Read more student research

Mean runoff of the river Thur in a present-day versus a pre-industrial climate

The thesis assesses the influence of anthropogenic greenhouse gas emissions on temperature, precipitation and runoff of the river Thur in Switzerland. To achieve this, thousands of Global Climate Model simulations with both present-day (A2000) and pre-industrial (A2000N) climate conditions are bias-corrected (from raw to corr; see Figure) and fed into a hydrological model for the river Thur. Higher mean runoff in winter and spring is formed in the present-day scenario (red colors) compared to the pre-industrial scenario (blue colors). This is mainly caused by the combination of increased precipitation and higher temperatures in winter, resulting in less snow accumulation in the present-day scenario. Therefore, more liquid water is available in the hydrological model, leading to enhanced runoff. In contrast, the present-day simulations exhibit lower runoff in summer and autumn than the pre-industrial simulations. This may be caused by the higher temperatures in the present-day scenario which enhance evapotranspiration and thus decrease the amount of water entering the hydrological model. These findings may support decision-makers in climate-related risk and adaptation management.
(Figure 5.7, page 41, from the master thesis by Regula Mülchi)

Read more student research

Weak points along the main rivers of the Bernese Oberland (Switzerland).

Weak points are locations along a river bed where the channel capacity is frequently exceeded during periods of high river discharges. Hence, weak points indicate potential focii for flood protection measures. Technically, weak points are defined as cross sections of the river bed where the bank-full discharge (channel capacity) has a flood recurrence period (HQ) between 30 and 100 years (red to yellow lines). Whereas specific river sections had been analyzed in practice so far, entire watersheds of the Bernese Oberland (Switzerland) were investigated in this thesis. The highest numbers of weak points (relative to the number of measured cross sections) are found for the Simme, Weisse Lütschine and the Hasli-Aare rivers. The good news is that frequently exceeded channel capacities (<= HQ30) are located along unpopulated areas. However, the protection aim of HQ100 along populated areas is not hold along parts of the Weisse Lütschine and the Simme, and neither along the Aare within the cities of Thun and Bern. These findings provide valuable information for both national and regional watershed administration.
(Figure 3.40, page 87, from the master thesis by Danielle Huser)

Read more student research

Seasonal total precipitation near Mt. Kenya, interpolated from gauge measurements during the Long Rains (March-April-May) in 1977.

The seasonal precipitation pattern in Central Kenya is highly influenced by the peculiar topography of the area, especially by Mt. Kenya (contour lines; bottom right quadrant). In this thesis, interpolations for seasonal and annual totals are obtained from rain gauge measurements (filled squares) using a combined approach of multiple linear regression with topographic predictors and ordinary kriging of residual precipitation. For the boreal spring rainy season, highest amounts of precipitation are found for the south-eastern slopes of the mountain.
Recent changes in agricultural practices and population growth have triggered conflicts between different water users, as economy and food security highly depend on rain-fed agriculture and river water discharge. In this context, the provided spatial information about rainfall can be used for water ressource allocations, for hydrological studies or to monitor trends over space and time.

(Figure 5.4, page 35, from the master thesis by Noemi Imfeld)

Read more student research

Cumulative influences of various drivers on the total CO2 emissions in the US from 1997 to 2013.

During the economic recession, between 2007 and 2009, the CO2 emissions dropped by approximately 10%, where the production structure as well as the consumption volume played the most important roles in said decline. The influence of the fuel mix was more or less continuously negative over the entire time period and sums up to the strongest negative influence on the CO2 emissions from 1997 to 2013. The fuel mix describes changes in energy use from coal to gas, from coal to renewables, from oil products to renewables, etc.

(Figure 2, page 51, from the master thesis by Marius Schneider)

Read more student research