A aplicação de gestão de riscos em eventos extremos como a seca é um fator primordial para redução de impactos sociais, além de auxiliar na utilização racional dos recursos naturais e financeiros. Na busca da eficiência da minimização dos efeitos da seca, vários países têm investido em estudos que utilizam índices para detectá-la, quantificá-la e monitorá-la. Entre os índices mais conhecidos têm-se o Índice Padronizado de Precipitação (SPI), que avalia a variação da precipitação em várias escalas de tempo, e o Índice de Aridez (IA), utilizado pela UNESCO para caracterização climática. Aliado a estes índices são desenvolvidos planos de gestão de riscos da seca. No Brasil a política de secas sempre foi voltada a gestão de crises, resultando num problema crônico de grandes impactos no Nordeste e, recentemente, também no Rio Grande do Sul (RS), como nos anos de 2004 e 2005. Nesta época, mais de 450 municípios decretaram situação de emergência devido à seca e houve a maior quebra de safra da história. Esta pesquisa objetivou avaliar a metodologia de gestão de secas utilizada pela Defesa Civil – RS, propondo a utilização de sistema de alerta baseado no SPI e IA e a adoção de um sistema de gestão de riscos fundamentado na avaliação, previsão e planejamento, composto de critérios objetivos na decretação de situação de emergência e estado de calamidade pública, além de diretrizes para um plano de ações para seca. Na comparação dos resultados dos índices de seca com as decretações de situação de emergência devido à seca no período de 1991 a 2005, o número de coincidências foi bastante baixo. Além disso, as regiões secas detectadas pelos índices de seca foram diferentes das regiões de maiores freqüências de decretações. Através desta análise pode-se concluir que a metodologia utilizada atualmente para gestão de secas não é eficiente. Um sistema de alerta para secas de curto prazo (agrícola) e longo prazo (hidrológica), baseado nos índices IA (mensal e trimestral) e SPI (anual), respectivamente, foi proposto e simulado. O IA identificou a região sul do RS como a mais propícia a períodos secos no estado, já o SPI identificou a região nordeste. Através dos resultados pôde-se concluir que os dois índices identificaram os mesmos períodos secos (meses e anos), porém, devido às regiões detectadas, o IA foi o que melhor representou a realidade climática do estado. A partir destes resultados foi proposta uma reformulação na metodologia da Defesa Civil considerando-se o sistema de alerta e análises de dados regionais para identificar as regiões mais vulneráveis a períodos secos. / The risk management in hazards like drought is very important to reduce social impacts, and to rationalize the natural and financial resources. To minimize the drought impacts, several countries take into account some indexes to its detection, quantification and evaluation. The Standardized Precipitation Index (SPI) is very applied around the world, and evaluates the variation of the precipitation in some timescale. Also, the Aridity Index (AI) is applied by UNESCO for climate characterization. These indexes should be a part of the plan of drought management risk. Usually, the drought policy of Brazil prioritizes the crisis management, with several negative impacts for decades in the Brazilian Northeast, and nowadays in Rio Grande do Sul. In 2004 and 2005, 450 municipals declared state of emergency by drought, with the biggest loss of harvest in local history. The research aim is to evaluate the drought management methodology of the Civil Defense of Rio Grande do Sul, and propose warning system based on SPI and AI as part of a risk management system to assess, forecast and plan actions against drought. This risk management system should include objective criteria to declare state of emergency and state of calamity, and guidelines for a drought plan. Comparing the results of drought indexes and declarations of state of emergence from 1991 to 2005, the number of coincidences was very low. Also, the dry regions detected by drought indexes were different of the regions with state of emergence. Thus, it can be concluded that the actual methodology for drought management is inefficient. It was proposed and tested a warning system for droughts of short term (agricultural), based on AI (monthly and quarterly), and droughts of long term (hydrological), based on SPI (annual). The AI identified the south of state as the drier, and the SPI identified the northeast. Both indexes identified the same periods of drought (months and years), but the AI, due the dry sites identified, best represents the local climatology. Thus, it was proposed a recast of the Civil Defense methodology for drought management, considering warning system and analysis of local data to identify the regions most vulnerable to dry periods.

A aplicação de gestão de riscos em eventos extremos como a seca é um fator primordial para redução de impactos sociais, além de auxiliar na utilização racional dos recursos naturais e financeiros. Na busca da eficiência da minimização dos efeitos da seca, vários países têm investido em estudos que utilizam índices para detectá-la, quantificá-la e monitorá-la. Entre os índices mais conhecidos têm-se o Índice Padronizado de Precipitação (SPI), que avalia a variação da precipitação em várias escalas de tempo, e o Índice de Aridez (IA), utilizado pela UNESCO para caracterização climática. Aliado a estes índices são desenvolvidos planos de gestão de riscos da seca. No Brasil a política de secas sempre foi voltada a gestão de crises, resultando num problema crônico de grandes impactos no Nordeste e, recentemente, também no Rio Grande do Sul (RS), como nos anos de 2004 e 2005. Nesta época, mais de 450 municípios decretaram situação de emergência devido à seca e houve a maior quebra de safra da história. Esta pesquisa objetivou avaliar a metodologia de gestão de secas utilizada pela Defesa Civil – RS, propondo a utilização de sistema de alerta baseado no SPI e IA e a adoção de um sistema de gestão de riscos fundamentado na avaliação, previsão e planejamento, composto de critérios objetivos na decretação de situação de emergência e estado de calamidade pública, além de diretrizes para um plano de ações para seca. Na comparação dos resultados dos índices de seca com as decretações de situação de emergência devido à seca no período de 1991 a 2005, o número de coincidências foi bastante baixo. Além disso, as regiões secas detectadas pelos índices de seca foram diferentes das regiões de maiores freqüências de decretações. Através desta análise pode-se concluir que a metodologia utilizada atualmente para gestão de secas não é eficiente. Um sistema de alerta para secas de curto prazo (agrícola) e longo prazo (hidrológica), baseado nos índices IA (mensal e trimestral) e SPI (anual), respectivamente, foi proposto e simulado. O IA identificou a região sul do RS como a mais propícia a períodos secos no estado, já o SPI identificou a região nordeste. Através dos resultados pôde-se concluir que os dois índices identificaram os mesmos períodos secos (meses e anos), porém, devido às regiões detectadas, o IA foi o que melhor representou a realidade climática do estado. A partir destes resultados foi proposta uma reformulação na metodologia da Defesa Civil considerando-se o sistema de alerta e análises de dados regionais para identificar as regiões mais vulneráveis a períodos secos. / The risk management in hazards like drought is very important to reduce social impacts, and to rationalize the natural and financial resources. To minimize the drought impacts, several countries take into account some indexes to its detection, quantification and evaluation. The Standardized Precipitation Index (SPI) is very applied around the world, and evaluates the variation of the precipitation in some timescale. Also, the Aridity Index (AI) is applied by UNESCO for climate characterization. These indexes should be a part of the plan of drought management risk. Usually, the drought policy of Brazil prioritizes the crisis management, with several negative impacts for decades in the Brazilian Northeast, and nowadays in Rio Grande do Sul. In 2004 and 2005, 450 municipals declared state of emergency by drought, with the biggest loss of harvest in local history. The research aim is to evaluate the drought management methodology of the Civil Defense of Rio Grande do Sul, and propose warning system based on SPI and AI as part of a risk management system to assess, forecast and plan actions against drought. This risk management system should include objective criteria to declare state of emergency and state of calamity, and guidelines for a drought plan. Comparing the results of drought indexes and declarations of state of emergence from 1991 to 2005, the number of coincidences was very low. Also, the dry regions detected by drought indexes were different of the regions with state of emergence. Thus, it can be concluded that the actual methodology for drought management is inefficient. It was proposed and tested a warning system for droughts of short term (agricultural), based on AI (monthly and quarterly), and droughts of long term (hydrological), based on SPI (annual). The AI identified the south of state as the drier, and the SPI identified the northeast. Both indexes identified the same periods of drought (months and years), but the AI, due the dry sites identified, best represents the local climatology. Thus, it was proposed a recast of the Civil Defense methodology for drought management, considering warning system and analysis of local data to identify the regions most vulnerable to dry periods.

A aplicação de gestão de riscos em eventos extremos como a seca é um fator primordial para redução de impactos sociais, além de auxiliar na utilização racional dos recursos naturais e financeiros. Na busca da eficiência da minimização dos efeitos da seca, vários países têm investido em estudos que utilizam índices para detectá-la, quantificá-la e monitorá-la. Entre os índices mais conhecidos têm-se o Índice Padronizado de Precipitação (SPI), que avalia a variação da precipitação em várias escalas de tempo, e o Índice de Aridez (IA), utilizado pela UNESCO para caracterização climática. Aliado a estes índices são desenvolvidos planos de gestão de riscos da seca. No Brasil a política de secas sempre foi voltada a gestão de crises, resultando num problema crônico de grandes impactos no Nordeste e, recentemente, também no Rio Grande do Sul (RS), como nos anos de 2004 e 2005. Nesta época, mais de 450 municípios decretaram situação de emergência devido à seca e houve a maior quebra de safra da história. Esta pesquisa objetivou avaliar a metodologia de gestão de secas utilizada pela Defesa Civil – RS, propondo a utilização de sistema de alerta baseado no SPI e IA e a adoção de um sistema de gestão de riscos fundamentado na avaliação, previsão e planejamento, composto de critérios objetivos na decretação de situação de emergência e estado de calamidade pública, além de diretrizes para um plano de ações para seca. Na comparação dos resultados dos índices de seca com as decretações de situação de emergência devido à seca no período de 1991 a 2005, o número de coincidências foi bastante baixo. Além disso, as regiões secas detectadas pelos índices de seca foram diferentes das regiões de maiores freqüências de decretações. Através desta análise pode-se concluir que a metodologia utilizada atualmente para gestão de secas não é eficiente. Um sistema de alerta para secas de curto prazo (agrícola) e longo prazo (hidrológica), baseado nos índices IA (mensal e trimestral) e SPI (anual), respectivamente, foi proposto e simulado. O IA identificou a região sul do RS como a mais propícia a períodos secos no estado, já o SPI identificou a região nordeste. Através dos resultados pôde-se concluir que os dois índices identificaram os mesmos períodos secos (meses e anos), porém, devido às regiões detectadas, o IA foi o que melhor representou a realidade climática do estado. A partir destes resultados foi proposta uma reformulação na metodologia da Defesa Civil considerando-se o sistema de alerta e análises de dados regionais para identificar as regiões mais vulneráveis a períodos secos. / The risk management in hazards like drought is very important to reduce social impacts, and to rationalize the natural and financial resources. To minimize the drought impacts, several countries take into account some indexes to its detection, quantification and evaluation. The Standardized Precipitation Index (SPI) is very applied around the world, and evaluates the variation of the precipitation in some timescale. Also, the Aridity Index (AI) is applied by UNESCO for climate characterization. These indexes should be a part of the plan of drought management risk. Usually, the drought policy of Brazil prioritizes the crisis management, with several negative impacts for decades in the Brazilian Northeast, and nowadays in Rio Grande do Sul. In 2004 and 2005, 450 municipals declared state of emergency by drought, with the biggest loss of harvest in local history. The research aim is to evaluate the drought management methodology of the Civil Defense of Rio Grande do Sul, and propose warning system based on SPI and AI as part of a risk management system to assess, forecast and plan actions against drought. This risk management system should include objective criteria to declare state of emergency and state of calamity, and guidelines for a drought plan. Comparing the results of drought indexes and declarations of state of emergence from 1991 to 2005, the number of coincidences was very low. Also, the dry regions detected by drought indexes were different of the regions with state of emergence. Thus, it can be concluded that the actual methodology for drought management is inefficient. It was proposed and tested a warning system for droughts of short term (agricultural), based on AI (monthly and quarterly), and droughts of long term (hydrological), based on SPI (annual). The AI identified the south of state as the drier, and the SPI identified the northeast. Both indexes identified the same periods of drought (months and years), but the AI, due the dry sites identified, best represents the local climatology. Thus, it was proposed a recast of the Civil Defense methodology for drought management, considering warning system and analysis of local data to identify the regions most vulnerable to dry periods.

<p>Predmet ovog istraživanja je suša i model osiguranja useva od rizika suše na teritoriji Vojvodine. Na osnovu detaljnog istraživanja vremenskih klimatoloških uslova na ispitivanom području, cilj ove disertacije je da se identifikuju meteorološki parametari i indikatori pogodni za indeksno osiguranje, potom da se identifikovani indeksi (padavine, SPI i SPEI indeksi suše) testiraju i odredi koji od njih najviše pogoduje osiguranju useva od rizika suše u Vojvodini.</p> / <p>The subject of this research is drought phenomena and crop insurance model for managing drought risk, on the territory of Vojvodina Province. After detailed research of weather and climate conditions of research area, themain purpose of this dissertation is to identify meteorological parameters and indicators suitable for index based insurance. Then,to test them (rainfall, SPI and SPEI drought indices) in order to determine index that is most suitable for crop insurance in Vojvodina.</p>

I Sverige finns det en brist på studier angående hydrologisk torka trots att det existerar problem med torka idag. Hydrologisk torka kan ha allvarliga konsekvenser på både naturen och samhället när det kommer till vattentillgång, växt -och djurliv och jordbruk. Av den anledningen är det viktigt att studier görs som undersöker allvarligheten i den hydrologiska torkan i Sverige för att få en bättre förståelse. I den här studien användes ett månadsvis Q95 värde som ett tröskelvärde med ett minimum av fem dagar i följd under tröskelvärdet för att definiera hydrologisk torka. Metoden applicerades på fem avrinningsområden in Sverige med data som sträckte sig mellan 1961- 2010. Resultatet från studien visade på att hydrologisk torka var speciellt framträdande under v issa år. Dessa år verkade vara kopplade till varandra under två till tre år i följd. De visade även ofta liknande månader och antal dagar under tröskelvärdet. Andra studier gjorda över de Nordiska länderna visade på liknande resultat. Metoden överensstämde även till en stor del av historisk torka i Sverige. För att kunna ge en större och komplett bild av hydrologisk torka diskuterades några möjliga metoder. Nederbörd, snö, strömflöde, evapotranspiration och grundvatten skulle behöva räknas med för en mer pr ecis studie. Standardiserade index kan täcka de mesta av de olika delarna, men för att få mera specifika förlustvärden så skulle även en tröskelnivå metod behöva implementeras i studien / In Sweden there is a lack of studies on the topic of hydrological drought even though it exist present problems of drought. Hydrological drought can have severe effects on both nature and society regarding water supply, animal life and agriculture. It is important to investigate the severity of hydrological drought in Sweden to get a better understanding of this phenomenon and its affects. To define hydrological drought this study used a Q95 monthly threshold with a minimum of 5 consecutive days below the threshold. This method was used on five catchments in Sweden with data ranging from 1961 -2010. The result from the study showed that hydrological drought was very prominent in some year s. These years seemed to be often linked together in two to three consecutive years. They often had similar amount of days and months below the threshold. Other studies over the Nordic countries showed similar results. The method also gave a result that to a certain degree showed droughts that coincided with historical records of drought in Sweden. This gave a positive feedback of the index accuracy. To get a broader picture of how hydrological drought propagates in Sweden some possible choices were discuss ed. Precipitation, snow, streamflow, evapotranspiration and groundwater would need to be covered for a more precise study. Standardized indices have most of spectrum covered, but it would be suggested to implement the threshold level method as well to get accurate deficits.

Meteorological drought indices are commonly calculated from climatic stations that have long-term historical data and then converted to a regular grid using spatial interpolation methods. The gridded drought indices are mapped to aid decision making by policy makers and the general public. This study analyzes the spatial performance of interpolation methods for meteorological drought indices in the United States based on data from the Co-operative Observer Network (COOP) and United States Historical Climatology Network (USHCN) for different months, climatic regions and years. An error analysis was performed using cross-validation and the results were compared for the 9 climate regions that comprise the United States. Errors are generally higher in regions and months dominated by convective precipitation. Errors are also higher in regions like the western United States that are dominated by mountainous terrain. Higher errors are consistently observed in the southeastern U.S. especially in Florida. Interpolation errors are generally higher in the summer than winter. The accuracy of different drought indices was also compared. The Standardized Precipitation and Evapotranspiration Index (SPEI) tends to have lower errors than Standardized Precipitation Index (SPI) in seasons with significant convective precipitation. This is likely because SPEI uses both precipitation and temperature data in its calculation, whereas SPI is based solely on precipitation. There are also variations in interpolation accuracy based on the network that is used. In general, COOP is more accurate than USHCN because the COOP network has a higher density of stations. USHCN is a subset of the COOP network that is comprised of high quality stations that have a long and complete record. However the difference in accuracy is not as significant as the difference in spatial density between the two networks. For multiscalar SPI, USHCN performs better than COOP because the stations tend to have a longer record. The ordinary kriging method (with optimal function fitting) performed better than Inverse Distance Weighted (IDW) methods (power parameters 2.0 and 2.5) in all cases and therefore it is recommended for interpolating drought indices. However, ordinary kriging only provided a statistically significant improvement in accuracy for the Palmer Drought Severity Index (PDSI) with the COOP network. Therefore it can be concluded that IDW is a reasonable method for interpolating drought indices, but optimal ordinary kriging provides some improvement in accuracy. The most significant factor affecting the spatial accuracy of drought indices is seasonality (precipitation climatology) and this holds true for almost all the regions of U.S. for 1-month SPI and SPEI. The high-quality USHCN network gives better interpolation accuracy with 6-, 9- and 12-month SPI and variation in errors amongst the different SPI time scales is minimal. The difference between networks is also significant for PDSI. Although the absolute magnitude of the differences between interpolation with COOP and USHCN are small, the accuracy of interpolation with COOP is much more spatially variable than with USHCN.

This study investigated the relationship between hydrological, and to some extent, meteorological droughts, and meteorological variables and catchment characteristics in 235 Swedish catchments between 1983 and 2013. This was done in order to investigate what factors affect the drought sensitivity in Swedish catchments and to map the occurrence of droughts in Sweden between 1983 and 2013. There have been studies about which meteorological phenomena and catchment characteristics that promote hydrological droughts, but for Sweden this is relatively unexplored. To investigate droughts during the study period three indices were used: the Standardized Precipitation Index (SPI), which is an index for meteorological droughts, the Standardized Streamflow Index (SSI), which predicts hydrological droughts and a threshold index for streamflow droughts. These indices were used to identify the number of drought events and the total number of drought days. For the majority of the 235 Swedish catchments there were no significant trends for the number of drought events or the total number of drought days during the 30-year period. The SPI and the SSI were found to correlate best in time when adding a one-month lag period to the SSI time series. The correlations between the indices and the meteorological variables and the catchments properties varied depending on how the catchments were grouped according to latitude or elevation. For example, the number of drought events was positively correlated to the mean elevation of the catchments in north and central Sweden when using the SSI while there were no significant correlations with elevation in southern Sweden. Another example is that it was almost only in northern Sweden where significant correlations between the percentage of bedrock and drought characteristics were identified. The percentage of bedrock can be used as an indication for how much groundwater a catchment can store. The correlations also look different for the different indices. For example, when looking at all catchments together the number of drought events identified with the SPI was negatively correlated to latitude and mean elevation while the number of drought events identified with the SSI was positively correlated to the same variables. For further research into this topic it would be wise to study winter and summer droughts separately to better identify which are the driving variables. / I den här studien undersöktes sambanden mellan hydrologiska, och till viss del meteorologiska, torrperioder och bakomliggande meteorologiska drivvariabler och avrinningsområdesegenskaper i 235 svenska avrinningsområden mellan 1983 och 2013. Detta gjordes i syfte att undersöka vilka faktorer som påverkar känsligheten för torka i svenska avrinningsområden och för att kartlägga förekomsten av torrperioder i Sverige mellan 1983 och 2013. Internationellt finns det studier på vilka meteorologiska fenomen och egenskaper hos avrinningsområden som leder till risk för fler torrperioder, men för Sverige är det ett relativt outforskat område. För att undersöka torrperioder under den aktuella perioden användes tre index: Standardized Precipitation Index (SPI), vilket är ett index för meteorologiska torrperioder, Standardized Streamflow Index (SSI), som används för hydrologiska torrperioder och ett tröskelvärdes-index för att identifiera hydrologisk torka. Indexen användes för att identifiera antalet torrperioder och totala antalet dagar med torka under studieperioden. För majoriteten av de 235 avrinningsområdena gick det inte att se några signifikanta trender för antalet torrperioder eller totala antalet dagar med torka under perioden 1983-2013. SPI och SSI korrelerade bäst med varandra över tiden när SSI-tidsserien försköts med en månad. Korrelationerna mellan torrperioderna identifierade med de olika indexen och de meteorologiska variablerna och avrinningsområdesegenskaperna varierade beroende på hur avrinningsområdena grupperades efter latitud eller medelhöjd. Till exempel, i norra och centrala Sverige korrelerade antalet torrperioder för SSI positivt med medelhöjden medan det i södra Sverige inte fanns några signifikanta korrelationer. Ett annat exempel är att det nästan bara var i norra Sverige som det fanns korrelationer mellan procenten berggrund och de identifierade torrperiodsegenskaperna. Procenten berggrund i jordlagret kan användas som en indikation på hur mycket grundvatten som kan lagars i ett avrinningsområde. Korrelationerna skiljde sig också åt för de olika indexen. Till exempel, sett över alla avrinningsområden så var antalet torrperioder beräknat med SPI negativt korrelerade med latitud och medelhöjd medan antalet torrperioder beräknat med SSI var positivt korrelerade med dessa egenskaper. För vidare forskning inom detta område rekommenderas att titta separat på vinter- och sommartorkor för att bättre kunna identifiera potentiella drivvariabler.

Floods and droughts are frequent in Central America and cause large social, economic and environmental impacts. A crucial step in disaster risk reduction is to have a good understanding of the causing mechanisms of extreme events and their spatio-temporal characteristics. For this, a key aspect is access to a dense network of long and good-quality hydro-meteorological data. Unfortunately, such ideal data are sparse or non-existent in Central America. In addition, the existing methods for hydro-climatic studies need to be revised and/or improved to find the most suitable for the region’s climate, geography and hydro-climatic data situation. This work has the ultimate goal to support the reduction of risks associated with hydro-climatic-induced disasters in Central America. This was sought by developing ways to reduce data-related uncertainties and by improving the available methods to study and understand hydro-climatic variability processes. In terms of data-uncertainty reduction, this thesis includes the development of a high resolution air temperature dataset and a methodology to reduce uncertainties in a hydrological model at ungauged basins. The dataset was able to capture the spatial patterns with a detail not available with existing datasets. The methodology significantly reduced uncertainties in an assumed-to-be ungauged catchment. In terms of methodological improvements, this thesis includes an assessment of the most suitable combination of (available) meteorological datasets and drought indices to characterise droughts in Central America. In addition, a methodology was developed to analyse drought propagation in a tropical catchment, in an automated, objective way. Results from the assessment and the drought propagation analysis contributed with improving the understanding of drought patterns and generating processes in the region. Finally, a methodology was proposed for assessing changes in both hydrological extremes in a consistent way. This contrasts with most commonly used frameworks that study each extreme individually. The method provides important characteristics (frequency, duration and magnitude), information that can be useful for decisions within risk reduction and water management. The results presented in this thesis are a contribution, in terms of hydro-climatic data and assessment methods, for supporting risk reduction of disasters related with hydro-climatic extremes in Central America. / Översvämningar och torka inträffar ofta i Mellanamerika och orsakar stora skador på samhälle, ekonomi och miljö. En kritisk del av riskreduceringen är förståelsen av mekanismerna bakom extremhändelserna, och deras rumsliga och tidskarakteristik. En nyckelfaktor är tillgång till långa tidsserier av rumsligt täckande hydrometeorologiska data av bra kvalitet. I Mellanamerika är sådana ideala data tyvärr sällsynta eller saknas helt. Dessutom behöver befintliga metoder för hydro-klimatisk analys revideras och/eller förbättras för att identifiera de mest lämpade metoderna för regionens klimat, geografi och situationen vad gäller hydrologiska och meteorologiska data. Det övergripande syftet med denna avhandling har varit att stödja arbetet med riskreducering i Mellanamerika vid hydrologiska extremhändelser som sätts igång av extrema väderhändelser. För att bidra till detta utvecklades metoder för att minska datarelaterade osäkerheter och för att förbättra tillgängliga metoder för att studera och förstå de processer som ligger bakom variabiliteten i hydrologi och klimat. Dataosäkerheten minskades genom utveckling av ett nytt dataset för lufttemperatur med hög rumslig upplösning och en metodik för att begränsa osäkerheten i modellberäknad vattenföring i ett område där det saknas observationer. Det nya datasetet kunde fånga rumsliga mönster på en detaljnivå som hittills inte varit möjlig. Metodiken möjliggjorde en klar minskning i osäkerheten hos vattenföringen i ett avrinningsområde som behandlades som om det saknade data. Avhandlingen innehåller också en metodik för att fastlägga den mest lämpade kombinationen av tillgängliga klimatdataset och torkindex för att karakterisera torka i Mellanamerika. Därutöver utvecklades en metod för att studera torkans fortplantning i ett tropiskt avrinningsområde på ett objektivt och automatiserat sätt. Slutligen föreslås en metod för att hantera förändringar av både översvämning och torka på ett konsistent sätt  som förenklar användningen av resultaten  för en beslutsfattare. Dessa metoder bedömdes användbara för att förbättra karakteriseringen och förståelsen av extrema hydrologiska händelser i Mellanamerika. Resultaten i denna avhandling ger bidrag till förståelsen av hydrologiska och klimatextremer genom förbättrade data och analysmetoder som i förlängningen kommer att stödja riskreduceringsarbetet i Mellanamerika. / Las sequías e inundaciones son frecuentes en Centroamérica y causan grandes problemas sociales, económicos y ambientales. Un aspecto crucial en la reducción del riesgo consiste en entender los mecanismos que causan dichos eventos, y sus características espacio-temporales. Para lograr esto es necesario tener acceso a una red de datos hidro-meterológicos densa, con series largas, y de buena calidad. Desafortunadamente, este no es el caso en Centroamérica. Además, los métodos para hacer estudios hidro-climáticos requieren ser evaluados y/o mejorados para asegurar su aplicabilidad en la región (su clima, su geografía y los datos disponibles). Este trabajo tiene como meta apoyar la reducción del riesgo de desastres asociados a eventos hidro-meteorológicos extremos en Centroamérica. Esto se consigue a partir de la reducción de incertidumbres asociadas a los datos, y de la mejora de métodos para el estudio de la variabilidad hidro-climática. Para reducir la incertidumbre de los datos, este trabajo incluye el desarrollo de una base de datos de temperatura de alta resolución y el desarrollo de una metodología para reducir las incertidumbres en datos simulados de caudal. Con la nueva base de datos se logra reconocer patrones espaciales a un nivel de detalle no antes captado por otras bases de datos. Por otro lado, la metodología redujo significativamente las incertidumbres de los datos simulados de caudal. En cuanto a métodos, esta tesis incluye una evaluación para encontrar la mejor combinación de índices de sequía y base de datos para la caracterización de sequías en la región. Además, se desarrolló una metodología para analizar la propagación de la sequía en una cuenca tropical, de una manera objetiva y automatizada. Los resultados de estos dos pasos ayudaron a mejorar la comprensión de los patrones y los mecanismos de generación de las sequías. Finalmente, se incluyó un método para evaluar los cambios en los patrones de sequías e inundaciones de una manera consistente, y no de manera individual como usualmente se ha hecho. Así fue posible obtener la frecuencia, duración y magnitud en ambos extremos hidrológicos. Esta información podría constituir una herramienta  útil para el manejo del riesgo y del recurso hídrico.

PhDH / Department of Hydrology and Water Resources / Demand for water resources has been on the increase and is compounded by population growth and related development demands. Thus, numerous sectors are affected by water scarcity and therefore effective management of drought-induced water deficit is of importance. Luvuvhu River Catchment (LRC), a tributary of the Limpopo River Basin in South Africa has witnessed an increasing frequency of drought events over the recent decades. Drought impacts negatively on communities’ livelihoods, development, economy, water resources, and agricultural yields. Drought assessment in terms of frequency and severity using Drought Indices (DI) in different parts of the world has been reported. However, the forecasting and prediction component which is significant in drought preparedness and setting up early warning systems is still inadequate in several regions of the world. This study aimed at characterising, assessing, and predicting drought conditions using DI as a drought quantifying parameter in the LRC. This was achieved through the application of hybrid statistical and machine learning models including predictions via a combination of hybrid models. Rainfall and temperature data were obtained from South African Weather Service, evapotranspiration, streamflow, and reservoir storage data were obtained from the Department of Water and Sanitation while root zone soil moisture data was derived from the NASA earth data Giovanni repository. The Standardised Precipitation Index (SPI), Standardised Precipitation Evapotranspiration Index (SPEI), Standardised Streamflow Index (SSI), and Nonlinear Aggregated Drought Index (NADI) were selected to assess and characterise drought conditions in the LRC. SPI is precipitation based, SPEI is precipitation and evapotranspiration based, SSI is based on streamflow while NADI is a multivariate index based on rainfall, potential evapotranspiration, streamflow, and storage reservoir volume. All indices detected major historical drought events that have occurred and reported over the study area, although the precipitation based indices were the only indices that categorised the 1991/1992 drought as extreme at 6- and 12- month timescales while the streamflow index and multivariate NADI underestimated the event. The most recent 2014/16 drought was also categorised to be extreme by the standardised indices. The study found that the multivariate index underestimates most historical drought events in the catchment. The indices further showed that the most prevalent drought events in the LRC were mild drought. Extreme drought events were the least found at 6.42%, 1.08%, 1.56%, and 4.4% for SPI, SPEI, SSI, and NADI, respectively. Standardised indices and NADI showed negative trends and positive upward trends, respectively. The positive trend showed by NADI depicts a decreased drought severity over the study period. Drought events were characterised based on duration, intensity, severity, and frequency of drought events for each decade of the 30 years considered in this study i.e. between 1986 – 1996, 1996 – 2006, 2006 – 2016. This was done to get finer details of how drought characteristics behaved at a 10-year interval over the study period. An increased drought duration was observed between 1986 - 1996 while the shortest duration was observed between 1996 - 2006 followed by 2006 - 2016. NADI showed an overall lowest catchment duration at 1- month timescale compared to the standardised indices. The relationship between drought severity and duration revealed a strong linear relationship across all indices at all timescales (i.e. an R2 of between 0.6353 and 0.9714, 0.6353 and 0.973, 0.2725 and 0.976 at 1-, 6- and 12- month timescales, respectively). In assessing the overall utilisation of an index, the five decision criteria (robustness, tractability, transparency, sophistication, and extendibility) were assigned a raw score of between one and five. The sum of the weighted scores (i.e. raw scores multiplied by the relative importance factor) was the total for each index. SPEI ranked the highest with a total weight score of 129 followed by the SSI with a score of 125 and then the SPI with a score of 106 while NADI scored the lowest with a weight of 84. Since SPEI ranked the highest of all the four indices evaluated, it is regarded as an index that best describes drought conditions in the LRC and was therefore used in drought prediction. Statistical (GAM-Generalised Additive Models) and machine learning (LSTM-Long Short Term Memory) based techniques were used for drought prediction. The dependent variables were decomposed using Ensemble Empirical Mode Decomposition (EEMD). Model inputs were determined using the gradient boosting, and all variables showing some relative off importance were considered to influence the target values. Rain, temperature, non-linear trend, SPEI at lag1, and 2 were found to be important in predicting SPEI and the IMFs (Intrinsic Mode Functions) at 1, 6- and 12- month timescales. Seven models were applied based on the different learning techniques using the SPEI time series at all timescales. Prediction combinations of GAM performed better at 1- and 6- month timescales while at 12- month, an undecomposed GAM outperformed the decomposition and the combination of predictions with a correlation coefficient of 0.9591. The study also found that the correlation between target values, LSTM, and LSTM-fQRA was the same at 0.9997 at 1- month and 0.9996 at 6- and 12- month timescales. Further statistical evaluations showed that LSTM-fQRA was the better model compared to an undecomposed LSTM (i.e. RMSE of 0.0199 for LSTM-fQRA over 0.0241 for LSTM). The best performing GAM and LSTM based models were used to conduct uncertainty analysis, which was based on the prediction interval. The PICP and PINAW results indicated that LSTM-fQRA was the best model to predict SPEI timeseries at all timescales. The conclusions drawn from drought predictions conducted in this study are that machine learning neural networks are better suited to predict drought conditions in the LRC, while for improved model accuracy, time series decomposition and prediction combinations are also implementable. The applied hybrid machine learning models can be used for operational drought forecasting and further be incorporated into existing early warning systems for drought risk assessment and management in the LRC for better water resources management. Keywords: Decomposition, drought, drought indices, early warning system, frequency, machine learning, prediction intervals, severity, water resources. / NRF