The Quasi-Biennial Oscillation (QBO) is the strongest phenomena influencing the stratopheric (~15-50 km height) circulation over the equator. QBO has two phases of downward propagating easterly and westerly winds, which has a total period of approximately 28 months and the phase is defined by the wind direction between the airpressure 25-50 hPa, which is roughly at a height of 30 km. QBO is induced by atmospheric gravity waves originating from the troposphere (~0-15 km height) and are generated by a plethora of sources, such as tropical convection and wind shear. The winds propagate downward at about 1 km per month through the stratosphere until reaching the tropopause (~15 km height) where they dissipate. The wind speed is at its maximum in the middle of the phase, where the wind shear is at its lowest, and the easterly winds can grow up to 30 m/s whilst the westerly winds reach roughly 15 m/s. Although the QBO is an equatorial phenomena it has a poleward component radiating its signal from the tropics to the higher latitudes where it affects other circulations such as the stratospheric polar vortex on the northern hemisphere (NH). The polar vortex consists of westerly winds around the polar region and is a major influence on the winter climate on the NH and thereby allows the QBO to indirectly affect the tropospheric climate through it. The easterly QBO disturbs and weakens the polar vortex, which results in warm subtropical air penetrating the vortex and warming the Arctic region whereas the polar air is released southward creating a colder winter on the NH. The westerly QBO on the other hand enhances the polar vortex and contains the cool polar air over the Arctic, which results in a milder winter. The correlation between QBO and El Niño Southern Oscillation (ENSO) as well as the tropical cyclones (TC) has either changed (ENSO) or completely disappeared (TC). The ENSO-QBO correlation depends on which phase of ENSO coincide with which phase of QBO, where El Niño coinciding with easterly QBO and La Niña coinciding with westerly QBO results in wind anomalies in the NH stratosphere. If the opposite combination takes place the wind anomalies will instead be situated in the subtropical troposphere, displacing the subtropical jet poleward. To what extent these stratospheric winds exert their influence is to some degree still uncertain, but that they have an effect on the tropospheric climate is unbeknownst to no one. / Cirkulationen i den ekvatoriella stratosfären (ca 15-50 km höjd) domineras av Quasi-Biennial Oscillation (QBO), ett zonalt (parallellt ekvatorn) vindfenomen med två faser bestående av östliga respektive västliga vindar och en period på ca 28 månader. Fasen definieras mellan lufttrycken 25-50 hPa, vilket representerar en höjd på ca 30 km. Drivkraften bakom QBO är ett brett spektrum av atmosfäriska gravitationsvågor som skapas genom bland annat den tropiska konvektionen, vindskjuvning och frontsystem. Vindarna propagerar vertikalt nedåt genom stratosfären med ungefär 1 km per månad tills de når tropopausen (ca 15 km) där vindarna försvagas kraftigt till ett zonalt medelvärde på 0 m/s. Vindhastigheten under östlig QBO uppgår i ca 30 m/s medan västlig QBO uppgår i ca 15 m/s, och är maximal i höga stratosfären samt i mitten av faserna där vindskjuvningen är minimal. QBO sprider sig meridionalt (nord-syd) från tropikerna till högre breddgrader genom stratosfären där andra fenomen som den stratosfäriska polarvirveln kan påverkas på norra halvklotet (NH). Polarvirveln består av västliga vindar i stratosfären runt polarregionen och är en stor influens på vinterklimatet i framförallt Europa och Nordamerika. Genom polarvirveln kan QBO indirekt påverka klimatet i troposfären (ca 0-15 km), där den östliga fasen av QBO försvagar medan den västliga fasen av QBO förstärker polarvirveln. En försvagad polarvirvel innebär en varmare medeltemperatur på Arktis och att kallare polarluft söker sig söderut och orsakar kalla vintertemperaturer. Troposfäriska klimatfenomen som El Niño Southern Oscillation (ENSO) och tropiska cykloner (TC) har uppvisat ett samband till QBO, men sedan förändrats (ENSO) eller helt försvunnit (TC). ENSO-QBO korrelationen förändras beroende på vilken fas QBO respektive ENSO är i relativt varandra. Då El Niño sammanfaller med östliga QBO samt La Niña sammanfaller med västliga QBO uppstår vindanomalier vid höga latituder i NH:s stratosfär, medan vid omvända sambandet förflyttar sig vindanomalierna till subtropikerna i troposfären och kan där förskjuta den subtropiska jetströmmen norrut. Att de stratosfäriska vindarna påverkar troposfären är känt, men hur och till vilken grad är ännu inte uppenbart. På grund av den korta tidsperiod med kontinuerliga och tillförlitliga vindmätningar i stratosfären uppkommer flera hypotetiska effekter av QBO och dess påverkan på klimatet i troposfären.

Glacial environments in the Arctic are a much-studied topic as well as a field of research with strong influences regarding the current and future global climate of our planet. This report is focused on the meteorological conditions on Nordenskiöldbreen glacier from 2009-2015 and how they correlate with each other, the glacier surface and the surrounding terrain. With data gathered from an automatic weather station located at the centre of the glacier, a range of meteorological parameters is examined; wind direction and velocity, snow depth, cloud cover, incoming and reflected shortwave radiation, temperature deficit, albedo and drifting snow. Seasonal differences were discovered, especially for wind direction and velocity where winds from the northeast occurred more frequently in the winter, indicating katabatic winds, whereas winds from the west and southwest were more pronounced in the summer. The calculated temperature deficit shows that katabatic winds blow down-glacier under stably stratified conditions and are shown to increase in strength with increasing temperature deficit (atmospheric temperature minus surface temperature). The mean albedo at Nordenskiöldbreen during this period is within the expected limits, 0.8 for snow and 0.3 for ice and the cloud cover was 0.58. Additionally, it could be observed that the occurrence of dry, drifting snow is present in the winter season as snow depth shows pronounced drops during high-wind events in winter. Overall, it is concluded that most of the examined parameters correlate and need each other to function and act as mechanisms within the cryosphere and as such it is crucial for scientists to understand their connected relationships when attempting to study global climate changes.

Deep artificial neural networks is a type of machine learning which can be used to find and utilize patterns in data. One of their many applications is as method for regression analysis. In this thesis deep artificial neural networks were implemented in the application of estimating the error of surface temperature forecasts as produced by a numerical weather prediction model. An ability to estimate the error of forecasts is synonymous with the ability to reduce forecast errors as the estimated error can be offset from the actual forecast. Six years of forecast data from the period 2010--2015 produced by the European Centre for Medium-Range Weather Forecasts' (ECMWF) numerical weather prediction model together with data from fourteen meteorological observational stations were used to train and evaluate error-predicting deep neural networks. The neural networks were able to reduce the forecast errors for all the locations that were tested to a varying extent. The largest reduction in error was by 83.0\% of the original error or a 16.7\degcs decrease in the mean-square error. The performance of the neural networks' error reduction ability was compared with that of a contemporary Kalman filter as implemented by the Swedish Meteorological and Hydrological Institute (SMHI). It was shown that the neural network implementation had superior performance for six out of seven of the evaluated stations where the Kalman filter had marginally better performance at one station.

In the last 100 years, a rise in the global mean temperature has been noted, and projections show even higher temperatures in the future. The temperature rise can lead to changes in the weather patterns and therefore the thunderstorm activity in a future warmer climate has been investigated in this study. The future projections were made with an ensemble of 8 General Circulation Models downscaled with the regional climate model RCA4, developed at SMHI. Temperature and humidity data at four different levels in the atmosphere has been used to compute three different stability indices. Stability indices indicate potential for deep convection in the atmosphere, from which thunderstorms are developed. It was found that the projections show an increase in thunderstorm potential in a warmer future climate. In Sweden, the projections show an increase with about 15 more days with risk of thunderstorms at the end of the 21st century for the RCP4.5 scenario, corresponding to an increase of 40% in the south, and an even larger increase in the north. For the RCP 8.5 scenario, the projected change in days with risk of thunderstorms corresponds to an increase about 20 days, or about 60% more thunderstorm days in south of Sweden. In other parts of Europe, the increase is expected to be even larger, mainly in the mountain regions. It was also found that the thunderstorm season is projected to be extended in the future, with more days with risk of thunder in May and September. The increase in number of days with risk of thunderstorms is a result of the greater amount of water vapour that the atmosphere is able to hold in a warmer climate. Even if thunderstorms are projected to increase, other factors counteract, such as a decrease in the vertical temperature gradient and a decrease in the difference between moisture in the upper and the middle atmosphere. Yet, taken together the days with risk of thunderstorms are projected to become more frequent. / Under de senaste hundra åren har medeltemperaturen på jorden ökat med cirka 1°C, vilket har medfört förändringar i klimatet. Temperaturen kommer att fortsätta stiga på grund av den redan förhöjda halten växthusgaser i atmosfären, och om växthusgaser fortsätter släppas ut kan det förväntas bli ännu varmare. I och med att temperaturen fortsätter stiga är det mycket som pekar på att vädret i allmänhet kommer förändras, som till exempel förändrat mönster i åskoväder. I denna studie har risken för åska i ett framtida klimat studerats. Åska och konvektion bildas framförallt på grund av tre komponenter: instabilitet i atmosfären, fuktigheten i luften och en mekanism som får luften att lyftas från marken. En instabil luftmassa fås framförallt en varm sommardag när solen värmer marken, vilket medför att luften vid marken blir betydligt varmare än luften ovanför. Den andra faktorn är beroende av fuktigheten i luften, om luften är tillräckligt fuktig finns risk att större åskmoln kan bildas. De första två faktorerna kan beskrivas med vad som kallas stabilitetsindex. I denna studie beräknades risken för djup konvektion med hjälp utav stabilitetsindex. Temperatur-och fuktighetsdata från den regionala klimatmodellen RCA4, framställd på SMHI, användes för att beräkna dessa stabilitetsindex. Studien visar på att dagar med risk för åska förväntas öka i slutet av detta seklet med omkring 10-15 dagar per år över Sverige, med ännu fler dagar med risk för åska i södra Europa. En förhöjd åskrisk kan även förväntas vid bergskedjor så som svenska fjällen och Alperna. Den främsta anledningen till att åska förväntas bli vanligare är till följd av att temperaturstigningen möjliggör högre halt vattenånga i atmosfären, och därmed kommer fuktigheten i luften att öka. En längre åsksäsong har även noteras, med tidigare start i maj, och även förlängd i september.

Orographic Rossby waves are the main mechanism by which the jet streams meander aroundthe Earth and have possibly far-reaching impacts on weather and climate (chapter 1). Hence,they are of particular importance to study and this project should serve as a starting point inwhat to consider when trying to model these waves. For example, we have to account forpressure gradients, Coriolis effect, orography, potential vorticity conservation and also Earth’scurvature at this scale. These are covered in detail in ch. 2 and adapted to the Shallow WaterEquations. In addition, some entry-level numerical techniques for solving these equations arepresented throughout ch. 2.4 and then implemented for the global-scale Shallow WaterEquations with conserved potential vorticity in ch. 3. The model is validated to work for typicalshallow water flows in a bath tub and passes common tests like the Gaussian curve test (ch.4.1). However, when considering atmospheric flows (ch. 4.2) it becomes evident that ourmodel, as well as our numerical methods are lacking and cannot reproduce Rossby waves ina stable manner. Hence, a heavily modified version of Hogan’s model (Hogan, n.d) isemployed with a simplified numerical scheme. With these corrections, orographic Rossbywaves appear to naturally form at appropriate locations. However, they do not fully exhibit theexpected behaviours discussed in ch. 2.2. Even Hogan’s model appears to have severelimitations as waves propagate in the wrong direction. Hence, this study is not complete andwarrants further development in order to be useful. / Orografiska Rossby-vågor är den huvudsakliga mekanismen genom vilken jetströmmarnaslingrar runt jorden och kan ha en omfattande inverkan på väder och klimat (kapitel 1). Därförär de av särskild betydelse att studera och detta projekt bör fungera som en utgångspunkt förvad man måste överväga när man försöker modellera dessa vågor. Till exempel så måste vi tahänsyn till tryckgradienter, Coriolis-effekten, orografi, potentiell vorticitetsbevarande och ävenjordens krökning på denna skala. Dessa beskrivs i detalj i kap. 2 och anpassas tillrörelseekvationerna för grunt vatten (Saint-Venant-ekvationerna). Därefter presenteras någranumeriska tekniker på grundläggande nivå för att lösa dessa ekvationer i kap. 2.4, varvid desedan implementeras för de globala Saint-Venant-ekvationerna med bevarad potentiellvorticitet i kap 3. Modellen är validerad för typiska grunda vattenflöden i ett badkar ochpasserar vanliga numeriska tester så som Gauss-kurvtestet (kap. 4.1) och bore-testet. Mennär vi överväger atmosfäriska flöden (kap. 4.2) blir det tydligt att våra modeller och numeriskametoder är primitiva och inte kan reproducera Rossby-vågor på ett stabilt sätt. Därmed,modifierar vi Hogans modell (Hogan, n.d) för att passa vår modell vilket resulterar orografiskaRossby-vågor. Dock så är dessa förskjutna och stämmer inte riktigt överens med teorin i kap.2.2. Även Hogans modell visar sig ha allvarliga begränsningar då vågorna propagerar i felriktning. Därmed är denna studie ej komplett och kräver ytterligare utveckling för att varaanvändbar.

Långtidsprognoser beskriver gapet mellan väderprognoser och klimatmodeller som förutspår klimatförändringar p.g.a. den globala uppvärmningen. Långtidsprognostik förlitar sig på så kallade “källor av förutsägbarhet”. Dessa kan vara variabler som ändrar sig långsamt, som havsvattnets yttemperatur, eller variationsmönster såsom El Nino-Southern Oscillation. Många industrisektorer och delar av samhället som ´ idag använder väderprognoser i sina dagliga beslut, som t.ex. jordbruksindustrin, energiindustrin eller någon annan väderkänslig sektor, kan potentiellt dra nytta av träffsäkra och pålitliga långtidsprognoser. Bland de potentiella användarna finns förstås också privatpersoner. Syftet med denna studie är att introducera långtidsprognostik och att försöka utvärdera prestandan av ECMWFs välkända och etablerade säsongsmodell SEAS5, genom tidigare forskning och ECMWFs verifikationsfigurer som finns tillgängliga för allmänheten. Utvärderingen gjordes utifrån allmänhetens perspektiv och vad de skulle uppfatta som en “bra” prognos. Marknära temperatur och delvis nederbörd undersöktes för Europa under sommar och vinter, i termer av “skill” (träffsäkerhet) och “reliability” (pålitlighet, på så sätt att en händelse som förutspås med 60% sannolikhet också ska observeras i ungefär 60% av fallen). SEAS5 når inte riktigt upp i “bra” skill-nivåer för marknära temperatur, om “bra” motsvarar synoptiskt användbar prognos. Det finns dock stora skillnader mellan olika platser och säsonger, där “bra” skill framförallt märks i södra/sydöstra Europa på sommaren och delar av norra Europa på vintern. “Reliability” är generellt bra, även om det är svårt att avgöra hur den skiljer mellan olika platser. Nederbörd visar dock mycket dålig skill och låg “reliability” oavsett säsong eller plats. / Long range weather prediction describe weather forecasts with a range longer than 14 days, but shorter than climate prediction models that predict climate change due to global warming. Long range forecasting relies on sources of predictability that, for example, changes slowly such as sea surface temperatures, or varies in predictable patterns like the El Nino-Southern Oscillation. Many sectors of industry and society ´ that today use weather forecasting in their day-to-day decision making, such as agriculture, energy or any other weather sensitive sector, have potentially much to benefit from accurate and reliable long range forecasts. Among potential users is of course the general public. The purpose of this study is to give an introduction to long range weather prediction and attempt to evaluate the performance of ECMWF’s SEAS5 seasonal model, which is one of the most well known and established S2S models, by using earlier research and ECMWF’s publicly available verification charts. This was done from the public’s perspective of what would be considered a “good” forecast, mainly from near surface air temperature but also precipitation, during winter and summer in Europe, on the aspects of skill (accuracy) and reliability (in the sense that a probabilistic forecast of 60% for an event also is observed around 60% of the time). SEAS5 overall doesn’t quite reach “good” skill levels for near surface temperature, if “good” is defined as synoptically significant. The skill level varies significantly though, depending on region and season, with southern/southeastern Europe during summer and parts of northern Europe during winter being notable “good” situations. Reliability is generally good, even if it is difficult to know how reliability varies spatially. However, precipitation shows very little skill and low reliability, no matter the season or region.

This master thesis is divided into two parts, one pedagogical and one engineering. The purpose of the pedagogical part of this master thesis was to investigate how students learn during projects. At the Royal Institute of Technology in Stockholm, three larger student projects occurred where the students themselves developed a probe that was launched into the atmosphere. The supervisors of the projects wanted to find out how the students learn during the project. The thesis includes in-depth interviews with current and former university students. In order to compare and gain new perspectives on learning,the study also included interviews with high school students to identify their corresponding experiences of learning in their final projects in Swedish upper secondary school. The result from this study shows that the students learn through participating inactivities, collaboration and communication. Giving the students responsibility, a mutual goal and an important assignment makes them collaborate and learn from experience through reflection. The purpose of the engineering part of this master thesis was to investigate samples that were collected during the Arctic Summer Cloud Ocean Study in the summer of 2008. The samples were studied by using a scanning electron microscope. The results of the thesis are consistent with former studies on samples collected in Arctic. The images from the microscope showed microgels and how the gels assembled into larger particles, particles which can play a crucial role in the formation of clouds.

The K-distribution method is a fast approximative method used for calculating thermal infrared radiative transfer in the atmosphere, as opposed to the traditional Line-by-line method, which is precise, but very time-costly. Here we consider the atmosphere to consist of homogeneous and plane-parallel layers in local thermal equilibrium. This lets us use efficient upwards recursion for calculating the thermal infrared radiative transfer and ultimately the outgoing irradiance at the top of the atmosphere. Our specific implementation of the K-distribution method revolves around changing the integration space from the wavenumber domain to the g domain by employing Gauss-Legendre quadrature in two steps. The method is implemented in MATLAB and is shown to be several thousand times faster than the traditional Line-by-line method, with the relative error being only 3 % for the outgoing irradiance at the top of the atmosphere.

Då de flesta människor i Västafrika, framför allt de som bor innanför kustzonen, livnär sig på jordbruk så är pålitliga väderprognoser och säsongsförutsägelser ett viktigt hjälpmedel i det dagliga arbetet och planeringen. I den här rapporten har en litteraturstudie gjorts för att öka kunskapen om de komplexa, både lokala och storskaliga, väderfenomen som ger upphov till nederbörd i området.   Klimatet i Västafrika, ett område som mestadels täcks av regnskog eller savann, präglas av den västafrikanska monsunen som ger regnperiod under norra halvklotets sommar och torrperiod på vintern. Denna monsun visar på stor årlig variation när det gäller dess inledande faser, och prognoser som kan förutsäga dess början är nödvändigt när en jordbrukare ska planera säsongens verksamhet. Av många bidragande faktorer framgår det tydligt att ytvattentemperaturen i Guineabukten är en av de viktigaste parametrarna för monsunens startskede. Den intertropiska konvergenszonen, ITCZ, betraktas som monsunens nordligaste del och denna konvergenszon gör en plötsligt och relativt snabb förflyttning norrut, i fortsättningen benämnd som språnget, över ca 5 breddgrader, vilket av många ses som starten på regnperioden. En tillfällig tryckgradient som uppstår på grund av en tillfällig men skarp temperaturgradient är den bakomliggande orsaken till detta språng. Den i särklass viktigaste processen som ger upphov till regn i Västafrika är konvektion och även om mycket konvektiv nederbörd faller i samband med monsunen så uppstår det även många lokala och mesoskaliga konvektiva system inom monsunen, framförallt kopplat till ostliga vågor. / Since most people in West Africa, particularly those who live away from the shore, work within agriculture, reliable weather forecasts are important in the daily work. In this report, a literature study has been conducted to increase the knowledge about the complex and local weather phenomenon that causes rain.   West Africa is dominated by rainforest and savannah, and the climate is characterized by the West African monsoon which gives rise to a wet season in the northern hemisphere summer and a dry season during the winter. The monsoon shows wide annual variability in its initial stages and forecasts that can predict its beginning is necessary to the farmers who have to decide when and what to plant. Out of many contributing factors, the sea surface temperature seems to be one of the main parameters that have an impact on the starting stage of the monsoon. The inter tropical convergence zone, ITCZ, is regarded as the northernmost part of the monsoon and it makes a sudden and relatively quick jump over about 5 degrees latitude, which is widely regarded as the start of the rainy season. A temporary pressure gradient caused by differences in temperature adjacent to the ITCZ is the reason for this jump. The dominant process that causes rainfall in West Africa is convection. Although a lot of convective precipitation falls in connection with the northern part of the monsoon, convective systems can appear inside the monsoon flow, often linked to African easterly waves.

Spatial distribution of precipitation in complex terrain can be very uneven and there are many factors to consider when studying it. The goal of this study was such a problem; to analyse the difference in observed annual precipitation in the complex terrain of Seyðisfjörður, a fjord in East-Iceland. The study was carried out in three parts. First, it was investigated if there was a systematic difference between measurement methods using different instruments. Second, the effect of the complex terrain on the spatial distribution of precipitation was investigated, and lastly if this observed difference could be linked to any certain kind of weather regimes. To simplify the analysis, only liquid precipitation was included in the data set.  In Seyðisfjörður there are three divergently located precipitation gauges of three different types and set up. At the Icelandic Meteorological Office in Reykjavík the same type of gauges are co-located with the exact same set up as in Seyðisfjörður. A statistical analysis of those measurements showed that there is a systematic undercatch in tipping bucket gauges compared to weighted capacity and standard accumulation gauges. However, the difference is insignificant in size compared to the observed difference in the complex terrain. The complex terrain was found to highly affect the airflow and therefore the spatial distribution of precipitation, as it almost only rains in synoptic wind directions with an easterly component (measured at a mountain station). To connect events where there was a great difference in precipitation measurements to the synoptic weather situation over the North-Atlantic, a projection connecting the geostrophic and local winds in the fjord was made. It showed that great precipitation as well as when great differences are observed, during two kind of weather regimes. One where a low pressure is travelling along the North-Atlantic storm track to the Norwegian Sea. The other were a low pressure is stationed southwest or west of Iceland in the Irminger Sea and a high pressure stretching up over Scandinavia. Convective precipitation makes up for a minimal part of the total precipitation and is not linked to events with great observed difference. Events with considerable precipitation but little observed difference are fewer and smaller than the events with much great observed precipitation and differences.