Cette thèse présente des contributions à la modélisation multivariée des queues de distribution. Nous introduisons une nouvelle modélisation des probabilités de queue jointes d'une distribution multivariée avec des marges Pareto. Ce modèle est inspiré de celui de Wadsworth et Tawn (2013). Une nouvelle variation régulière multivariée non-standard de coefficient une fonction à deux variables est introduite, permettant de généraliser deux approches de modélisation respectivement proposées par Ramos et Ledford (2009)et Wadsworth et Tawn (2013). En nous appuyant sur cette modélisation nous proposons une nouvelle classe de modèles semi-paramétriques pour l'extrapolation multivariée selon des trajectoires couvrant tout le premier quadrant positif. Nous considérons aussi des modèles paramétriques construits grâce à une mesure non-négative satisfaisant une contrainte qui généralise celle de Ramos et Ledford (2009). Ces nouveaux modèles sont flexibles et conviennent tant pour les situations de dépendance que d'indépendance asymptotique. / This PhD thesis presents contributions to the modelling of multivariate extremevalues. We introduce a new tail model for multivariate distribution with Pareto margins. This model is inspired from the Wadsworth and Tawn (2013) one. A new non-standard multivariate regular variation with index equals to a function of two variables is thus introduced to generalize both modeling approaches proposedby Ramos and Ledford (2009) and Wadsworth and Tawn (2013), respectively. Building on this new approach we propose a new class of non-parametric models allowing multivariate extrapolation along trajectories covering the entire first positive quadrant. Similarly we consider parametric models built with a non-negative measure satisfying a constraint that generalizes the Ramos and Ledford (2009) one. These new models are flexible and valid in both situations of dependence or asymptotic independence.

Ce travail de thèse concerne la caractérisation de la structure spatio-temporelle des fortes précipitations dans la région Cévennes-Vivarais. La région est soumise à des événements de pluie catastrophiques dont la magnitude gouverne les conséquences à différentes échelles de temps et d'espace. La détermination de la probabilité d'occurrence des orages est problématique à cause du caractère extrême des ces événements, de leur dimension spatio-temporelle et du manque de données pluviométriques aux échelles d'intérêt. Nous proposons d'adopter des approches d'invariance d'échelles afin d'estimer la fréquence d'occurrence de ces événements. Ces approches permettent d'extrapoler la distribution de la pluie à haute résolution à partir de données d'intensité pluvieuse à plus faible résolution. La paramétrisation de ces modèles étant fortement dépendante de l'incertitude de la mesure, nous avons d'abord caractérisé l'erreur commise dans la mesure de la pluie par un réseau de pluviomètres à augets. Nous avons ensuite exploré le comportement des pluies extrêmes dans la région d'étude, identifiant les gammes d'invariance d'échelles des extrêmes. Dans cette gamme d'échelles, nous présentons un modèle régional Intensité-Durée-Fréquence qui prend en considération l'hétérogénéité spatiale des extrêmes dans la région. Étant donné que le réseau pluviométrique ne permet pas de détecter les propriétés d'invariance d'échelle spatiale des champs de pluie, nous avons adopté une méthode semi-empirique pour modéliser des intensités de pluie intégrés sur des surfaces données (pluie surfacique) sur la base du concept de la mise en échelle dynamique (« dynamic scaling »). Cette modélisation permet la construction d'un modèle régional Intensité-Durée-Fréquence-Surface. Enfin, nous avons appliqué ce modèle à la construction des diagrammes de sévérité pour trois événements marquants en région Cévennes-Vivarais, afin d'identifier les échelles spatio-temporelles critiques pour chaque événement. Grâce aux diagrammes de sévérité, nous avons pu évaluer, pour ces mêmes événements, la performance d'un modèle météorologique de méso-échelle. / The thesis is devoted to the characterization of the space-time structure of heavy rainfall events in the Cévennes-Vivarais area (France). The region is prone to catastrophic storms whose magnitude governs social and economic consequences at different space and time scales. The magnitude of an event cannot be univocally related to a probability of occurrence. The determination of the occurrence probability of storms is problematic because of their extreme character, of their complex space-time development and of the lack of rainfall data at the spatial and temporal scales of interest. We propose to adopt scale-invariant approaches in order to estimate the heavy rainfall frequency assessment. These approaches allow to extrapolate the high resolution rainfall distribution based on low resolution rainfall intensity data. The model estimation being heavily dependent of the data accuracy, the first step consists in the characterization of the error committed in the point and spatial rainfall estimated by tipping-bucket raingage networks. We then explore the extreme rainfall behavior in the region, detecting the range where extremes are scale-invariant. In this range, we present a regional Intensity-Duration-Frequency model for point rainfall maxima taking into account the heterogeneity of extremes in the region. We demonstrate that the rainfall network does not allow to detect scale-invariant properties of extreme rainfall fields, and then we adopt a semi-empirical method based on the concept of « dynamic scaling » to build regional Intensity-Duration-Area-Frequency curves. Finally, we apply this model for the determination of the severity diagrams for three significant storms in the Cévennes-Vivarais region, with the aim to identify the critical space-time scales of each event. Based on severity diagrams, we then evaluate, for the same events, the performances of a mesoscale meteorological model.

Submitted by Lucienne Costa (lucienneferreira@ufcg.edu.br) on 2018-08-13T17:25:26Z No. of bitstreams: 1 JORDANNA SOUSA DE MELO – DISSERTAÇÃO (PPGMET) 2017.pdf: 3551837 bytes, checksum: 84c4d559fd24d6944745d0099094fe06 (MD5) / Made available in DSpace on 2018-08-13T17:25:26Z (GMT). No. of bitstreams: 1 JORDANNA SOUSA DE MELO – DISSERTAÇÃO (PPGMET) 2017.pdf: 3551837 bytes, checksum: 84c4d559fd24d6944745d0099094fe06 (MD5) Previous issue date: 2017-12-15 / Com o intuito de identificar e analisar as ondas de calor e de frio que ocorrem no Rio Grande do Sul foi utilizado quatro diferentes procedimentos em 13 estações meteorológicas, espacialmente distribuídas no Estado, entre os anos de 1961 a 2010. Foram consideradas ondas de calor e frio os períodos de cinco ou mais dias consecutivos de anomalias positivas e negativas de temperaturas máximas e mínimas, respectivamente. Calculou-se as médias climatológicas de temperaturas máximas e mínimas anual, estacional e diárias, e em seguida o número de ondas de calor e de frio para cada localidade, anual e sazonal pelos métodos Índice da Organização Meteorológica Mundial (IOMM), Índice Diário (ID), Índice Sazonal (IS) e RClimdex. Ao comparar os métodos verificou-se que o IOMM foi o que detectou o maior número de eventos, os métodos ID e IS praticamente não detectam ondas de calor no verão e de frio no inverno e o RClimdex mostrou-se incapaz de detectar a variabilidade internual de ondas de calor e frio no Rio Grande do Sul. Com relação às configurações espaciais das médias climatológicas das temperaturas do ar máximas e mínimas durante os dias de ondas de calor e frio, respectivamente, assim como os próprios números de dias de ondas calor e frio observou-se um gradiente do litoral para o interior e de sudeste para noroeste. De certa forma estas configurações é uma resposta aos efeitos da continentalidade, latitude e relevo. Quanto à associação da influencia dos eventos El Niño e La Niña no número de ondas de calor e frio observou-se que, em média, ocorrem um número maior de ondas de calor nos anos de El Niño e de frio nos anos de La Niña, entretanto, esta relação não é estatisticamente significativa. Portanto, não é possível afirmar categoricamente que em anos de El Niño tem-se um maior número de ondas de calor e nos de La Niña de frio. / In order to identify and analyze the heat and cold waves that occur in Rio Grande do Sul four different procedures were used in 13 meteorological stations, spatially distributed in the State, between 1961 and 2010. Heat waves were considered and periods of five or more consecutive days of positive and negative anomalies of maximum and minimum temperatures, respectively. The annual, seasonal and daily maximum and minimum climatic averages were calculated, followed by the number of heat and cold waves for each locality, annual and seasonal, using the World Meteorological Organization (IOMM), Daily Index (ID), Seasonal Index (IS) and RClimdex. When comparing the methods it was verified that the IOMM was the one that detected the greatest number of events, the ID and IS methods practically did not detect heat waves in summer and cold in the winter and RClimdex was unable to detect the internal variability of heat and cold waves in Rio Grande do Sul. Regarding the spatial configurations of the climatological means of the maximum and minimum air temperatures during the days of heat and cold waves, respectively, as well as the number of days of heat waves and cold it was observed a gradient of the coast inland and from southeast to northwest. In some ways these settings are a response to the effects of continental, latitude and relief. As for the association of the influence of the El Niño and La Niña events on the number of heat and cold waves, it was observed that, on average, a larger number of heat waves occur in the El Niño and cold years in La Niña, however, this relationship is not statistically significant. Therefore, it is not possible to state categorically that in El Niño years there is a greater number of heat waves and that of La Niña of cold.

L’étude vise à caractériser les indicateurs relatifs aux évènements hydroclimatiques extrêmes dans le bassin-versant de l'Ouémé à l'exutoire de Bonou. Pour atteindre cet objectif, les données climatologiques (pluies journalières, températures journalières, évapotranspiration potentielle journalières), les données hydrométriques (débits journaliers), les données de réanalyses (pression atmosphériques, humidité relative, vitesse du vent, Indices de Mousson West Africain), la température de surface océanique sur la côte de Cotonou et les données cartographiques relatives aux occupations du sol des années 1986, 2000 et 2015 sont utilisées. L’analyse de la variabilité hydroclimatique a révélé que le secteur d’étude a connu sur la période de 1951-2015, deux ruptures de stationnarité (1968 et 1987) qui ont divisé la série en trois sous-périodes. La variabilité pluviométrique révèle qu’il une régression de 18 % (au sud), de 16 % (au centre) et de 15,8 % (au nord) entre la première sous-période (1951-1968) et la deuxième sous-période (1969-1987), alors que la dernière sous-période (1988-2015) est marquée par une reprise pluviométrique de 15 % (au sud), de 16,3 % (au centre) et de 15,7 % (au nord) par rapport à la sous-période d’avant. Ce qui n’est pas sans conséquences sur la dynamique hydrologique du secteur d’étude. Ainsi, il est observé une régression de 49,27 % des débits moyens annuels entre les sous-périodes 1951-1968 et 1969-1987 et une augmentation de 65 % entre les sous-périodes 1969-1987 et 1988-2015. Sur la dernière sous-période (1988-2015), les résultats révèlent aussi une augmentation des indicateurs des évènements hydroclimatiques extrêmes susceptibles d’occasionner les inondations à Bonou. Les pluies maximales journalières ont une occurrence de 2 ans, 5 ans, 10 ans pour les événements pluvieux forts ; 20 ans, pour les événements pluvieux très forts ; 50 ans et + pour les événements pluvieux extrêmement forts dans le secteur d’étude. Les débits maximaux en 24 heures augmentent en même temps que les périodes de retour. Les débits maximaux de 50ème quantile surviennent chaque deux ans, les débits maximaux correspondants aux évènements hydrologiquement forts, très forts, surviennent tous les 10 ans et les débits extrêmement forts ont une occurrence centennale. Les conditions atmosphériques, océaniques, de la surface continentale, la vitesse du vent, l’humidité relative, les facteurs géographiques sont autant de facteurs qui interagissent dans la répartition spatio-temporelle de la pluie du secteur d’étude. La pluviométrie du bassin est également influencée par le flux de mousson (nord secteur d’étude) auquel s’associent les lignes de grains de sud-est, responsables des pluies orageuses et d’averses qui génèrent des inondations dans le bassin. De même, le contexte climatique actuel et la dynamique de l’occupation des terres constituent des déterminants qui amplifient la manifestation des inondations dans le secteur d’étude. La recherche indique que le modèle conceptuel pluie-débit GR4J, surestime les débits observés en période de basse eau et les sous-estime en période de hautes eaux. Les critères d’efficacité et de performance NSE, RMSE et KGE, mis en évidence et calculés sur les débits de hautes eaux ont donné des résultats meilleurs en calage qu’en validation. Mieux les valeurs du KGE varient de 83 à 85 % en calage et de 56 à 68 % en validation, ce qui confère au modèle GR4J, l’efficacité, la performance à reproduire les débits extrêmes de crues dans le bassin. Le GR4J peut donc être servir comme outil d’aide à la décision pour l’actualisation des normes hydrologiques dans le secteur d’étude. Dans la perspective de mieux étudier les évènements climatiques extrêmes futurs, cette recherche a évalué des données de projection de sorties des modèles climatiques régionaux issues du Programme CORDEX africain et qui pourraient être utilisées pour les projections hydrologiques des travaux futurs dans le secteur d’étude. / This research aims to characterize the indicators relating to extreme hydroclimatic events in Ouémé Watershed at Bonou’s outlet. To achieve this objective, climatological data (daily rain, daily temperatures, potential daily evapotranspiration), hydrometric data (daily flows), reanalysis data (atmospheric pressure, relative humidity, wind speed, West African Monsoon Indices), the ocean surface temperature on the Cotonou coast and land use mapping data of 1986, 2000, 2015 are used. The analysis of the hydroclimatic variability revealed that during the 1951-2015 period, the study area is marked by two breaks stationarity (1968 and 1987) which divided the series into three sub-periods. The rainfall variability reveals a regression of 18 % (in the south), 16 % (in the center) and 15.8 % (in the north) between the first sub-period (1951-1968) and the second sub-period (1969-1987), while the last sub-period (1988-2015) is marked by a rainfall recovery of 15 % (in the south), 16.3 % (in the center) and 15.7 % (in north) relative to the previous sub-period (1969-1987). This is not without consequences on the hydrological dynamics of the study area. In this sense, it is noted that there is a regression of 49.27% of the average annual flows between the 1951-1968 and 1969-1987 sub-periods and a 65 % increase between the 1969-1987 and the 1969-1987 sub-periods. 1988-2015. In the last sub-period (1988-2015), the results also reveal an increase in the indicators of extreme hydroclimatic events likely to cause floods Ouémé Watershed at Bonou’s outlet. The results shows that daily maximum rainfall has an occurrence of 2 years, 5 years, 10 years for strong rainfall events; 20 years, for very heavy rain events, 50 years and over for extremely heavy rain events in the study area. The maximum flow rates in 24 hours at the sametime as the return periods. The maximum flow rates of 50th quantiles occur every two years (02 years), the maximum flows corresponding to the hydrologically strong events, very strong occur every 10 years and the extremely high flows have an occurrence centennial. The conditions of atmospheric, oceanic, continental surface, wind speed, relative humidity, and geographic factors are all elements that interact in the spatio-temporal distribution of rainfall in the study area. The Rainfall of the basin is also influenced by the monsoon flow (north of the study area), to which the southeastern grain lines are associated and are responsible for the stormy rains and showers that generate flooding in the area. In addition, the current climatic context and land-use dynamics are determinants that amplify the occurrence of floods in the study area. The reasearch indicates that the rain-flow conceptual model, of the Rural Engineering with four parameters and the daily time step (GR4J), overestimates the flows observed during the low water period and the sub-peaks. The efficiency and performance criteria NSE, RMSE and KGE, highlighted and calculated on high water flow rates, gave better results in calibration than in validation. Better KGE values range from 83 to 85 % calibration and from 56 to 68 % validation, which gives the GR4J model the efficiency and performance to reproduce extreme flows of floods in the study area. The GR4J can therefore be used as a decision to support tool for updating hydrological standards in the study area. In order to future study about extreme climatic events, this research assessed output projection data from regional climate models from the African CORDEX Program that could be used for hydrological projections in future work in Ouémé Watershed at Bonou’s outlet.

Thesis (Ph. D. (Statistics)) -- University of Limpopo, 2016. / Statistics of extremes has seen much growth both in theory and application since its early theoretical developments almost a century ago in the 1920s and its first major applications to real-life problems pioneered by Emil Gumbel in the early 1940s. Although the theory and applications of extreme value theory (EVT) have been extensively advanced and utilised in most developed countries,intermsofapplicationslittlehasbeendoneinmanydevelopingcountries in Africa despite the abundance of areas of applications and raw data in some ofthesecountries. Inhydrology,thechoiceoffloodfrequencyprobabilitydistributions for a particular site or region remains the subject of ongoing research. The work contained in this thesis is a contribution towards this area and it addresses this problem in one of the developing and economically challenged countries in Africa, Mozambique, in the lower Limpopo River basin (LLRB). The LLRB is a basin characterised by extreme natural hazards, alternating between extreme floods and severe droughts. ThisthesisisbasedonanextensiveapplicationofEVTtoextremefloodheights data in the LLRB of Mozambique at three sites: Chokwe, Combomune and Sicacate hydrometric stations. Two fundamental approaches of EVT, block maxima and peaks-over-threshold (POT), are used in this thesis. Recent theoretical results by Ferreira and de Haan (2015) have shown that despite its inefficiency due to data lost as a result of blocking, the block maxima approach is more efficient in a number of situations than the POT approach, and the two approaches are quite comparable for large sample sizes. A number of ii candidate distributions are investigated for their goodness-of-fit to the annual daily maximum flood heights in a block maxima realisation at each site. The findings reveal that the GEV distribution is the most appropriate distribution to apply in the LLRB and the distribution can be recommended as the likelihood function for regional and spatial extremes flood frequency analysis in the basin. The thesis addresses the issue of cumulative effects on daily flood heights through a comparative analysis of six annual maxima moving sums. The findings demonstrate that the six annual maxima time series models are notsignificantlydifferentbasedonthecharacteristicsconsideredinthisthesis. In an attempt to reduce uncertainties in the estimates, a Bayesian Markov chain Monte Carlo (MCMC) approach with a conjugate prior and a GEV likelihood function is used to model the tails of the extreme flood heights in the basin. The findings reveal that the addition of prior information in Bayesian MCMC substantially reduces uncertainties in the estimates and improves precision in the predicted extreme floods. The r largest order statistics models developed in this thesis are generally promising and the standard errors of the estimates of the parameters are substantially reduced. In order to account for climate change impact, nonstationary models are considered with the longterm trend and seasonal oscillation index (SOI) (a meteorological variable indicator) as covariates of the parameters of the GEV distribution and the generalised Pareto distribution (GPD). Among the major contributions of this thesis is a proposed procedure for the determination of the 8 days window period used in extracting independent r largest order values within the same year for the r largest order statistics approach. A summary of the key findings and contributions of this thesis are given in Chapter 9. Moreover, contributions by the study topic in each chapter are given at the end of each chapter. / DST-NRF Centre of Excellence in Mathematical and Statistical Sciences (CoEMaSS) of South Africa

京都大学 / 新制・課程博士 / 博士(工学) / 甲第23164号 / 工博第4808号 / 新制||工||1752(附属図書館) / 京都大学大学院工学研究科社会基盤工学専攻 / (主査)教授 立川 康人, 准教授 KIM SUNMIN, 教授 中北 英一 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DFAM

Thesis (M.Sc. (Statistics )) -- University of Limpopo, 2021 / Temperature extremes have a crucial impact on agricultural, economic, health and energy sectors due to the occurrence of climate extreme events such as heat waves and cold waves. Limpopo province is among the hottest provinces of South Africa and experiences little rainfall which affect the water availabil ity, food production and biodiversity. In the Limpopo province, temperature extremes are expected to become more frequent as a result of climate change. The aim of this study was to model temperature extremes in the Limpopo province of South Africa using extreme value theory (EVT). The stationarity of the data was tested using augmented Dickey-Fuller (ADF), Phillips-Peron (PP) and Kwiatkowski-Phillips-Schmit-Shin (KPSS). Four candidate parent distri butions: normal, log-normal, gamma and Weibull distributions, were fitted to the average monthly maximum and minimum daily temperatures. Prior to the selection of the parent distributions, the data set at each station was subjected to normality test using the Shapiro-Wilk (SW) and Jarque-Bera (JB) tests. The stationarity and normality tests revealed that the maximum and minimum temperature data series at all the stations are neither stationary nor normally distributed. Akaike information criterion (AIC) and Bayesian information cri terion (BIC) were used to select the best fitting distribution at a particular site. The findings revealed that both maximum and minimum temperatures series at all the stations belong to the Weibull domain of attraction. The findings from the Mann-Kendall (M-K) test and time series plots trend analyses showed that there is a monotonic downward and upward long-term trend in minimum and maximum temperature data, respectively. Two fundamental approaches of EVT, block maxima and peaks-over-threshold (POT), were used in this dissertation. The generalised extreme value (GEV), generalised Pareto (GP) and Poisson point process distributions were fitted to the data set for each station. In order to account for climate change impact, non-stationary models were considered with Seasonal Oscillation Index (SOI) as covariates of the parameters of the GEV distribution. The findings revealed that both the maximum and minimum temperature data can be modelled by the Weibull family of distribution. The EVT return level analysis findings of above 400C for maximum temperature suggests impending heat waves and droughts in the Limpopo province. The bivariate conditional extremes ap proach with a time-varying threshold was used. The findings revealed both significant positive and negative extremal dependence in some pairs of meteo rological stations. The findings of this study play an important role in revealing information useful to meteorologists, climatologists, agriculturalists and plan ners in the energy sector where temperature extremes play an important role. The scientific contribution of this study was to reduce the risk and impact of temperature extremes on agricultural, energy and health sectors in the Limpopo province. An understanding of temperature extremes will help gov ernment and other stakeholders to formulate mitigation strategies that will minimise the negative impact resulting from temperature extremes in the Limpopo province. Among the major contributions of the study was the use of a pe nalised cubic smoothing spline to perform a nonlinear detrending of the tem perature data, before fitting bivariate time-varying threshold excess models based on Laplace margins, to capture the climate change effects in the data. Future studies may consider exploring the use of extreme value copulas, as well as spatio-temporal dependence between temperature extremes using the conditional extremes model of Heffernan and Tawn (2004). / DST National Research Foundation (NRF)

No description available.

Weather and climate models applied with sufficiently fine mesh grids to enable a large part of atmospheric deep convection to be explicitly resolved have shown a significantly improved representation of local, short-duration and intense precipitation events compared to coarser scale models. In this thesis, two studies are presented aimed at exploring the dependence of horizontal resolution and of parameterization of convection on the simulation of precipitation. The first examined the ability of HARMONIE Climate (HCLIM) regional climate model to reproduce the recent climate in Europe with two different horizontal resolutions, 15 and 6.25 km. The latter is part of the ”grey-zone” resolution interval corresponding to approximately 3-10 km. Particular focus has been given to rainfall and its spatial and temporal variability and other characteristics, for example intensity distributions. The model configuration with the higher resolution is much better at simulating days of large accumulated precipitation amounts, most evident when the comparison is made against high-resolution observations. Otherwise, the two simulations show similar skill, including the representation of the spatial structure of individual rainfall areas of primarily convective origin. The results suggest a ”scale-awareness” in HCLIM, which supports a central feature of the model’s description of deep convection as it is designed to operate independently of the horizontal resolution. In the second study, summer season precipitation over the Alps region, as simulated by HCLIM at different resolutions, is investigated. Similar model configurations as in the previous study were used, but in addition a simulation at the ”convection-permitting” 2 km resolution has been made over Central Europe. The latter considerably increases the realism compared to the former regarding the distribution and intensities of precipitation, as well as other important characteristics including the duration of rain spells, particularly on sub-daily time scales and for extreme events. The simulations with cumulus parameterization active underestimate short-duration heavy rainfall, and rainspells with low peak intensities are too persistent. Furthermore, even though the 6.25 km simulation generally reduces the biases seen in the 15 km run, definitive conclusions of the benefit of ”grey-zone” resolution is difficult to establish in context of the increased requirement of computer resources for the higher-resolution simulation.

Revegetation projects in wetlands are challenging due to questions surrounding where to obtain plant materials and how hydrologic conditions, which are often unpredictable at restoration sites, may impact restoration success. We used a two-pronged approach to inform decisions on seed sourcing. Our study species, Schoenoplectus maritimus (alkali bulrush), is a widely distributed wetland plant. First, we investigated how genetic diversity was partitioned within and among populations of S. maritimus. We found five weakly differentiated populations and one distinct population. We found high levels of genetic diversity with the majority (92%) of diversity found within rather than among sites (8%). Also, the proportion of viable seed produced was surprisingly high within stands (mean = 0.64 ± 0.02) given the supposed prevalence of asexual reproduction in the species. Second, we conducted two studies to look at the influence of hydrology, population of origin, and genetic diversity of seeds on the productivity of S. maritimus. In a field survey we measured environmental variables and productivity within established S. maritimus stands. In a greenhouse experiment we determined how source population identity and the genetic diversity of seeds impacted emergence and productivity under different hydrologic conditions. We found that stands of S. maritimus differed in proportion of time with water present, mean water level, and soil conditions. Productivity also differed, with 3-fold differences in stem density and biomass among sites. In the greenhouse experiment, we found that water treatment impacted all productivity measures; source population impacted seedling emergence and biomass allocation; and, number of source populations impacted sensitivity to drought. Advice for future restoration projects includes (1) limiting translocation of seeds among populations to conserve historic lineages, (2) when it is necessary to translocate seeds, collect seeds from many parent plants within populations that are in close geographic proximity to the restoration site, and (3) water level management is extremely important at all life stages of S. maritimus and should be an important consideration in wetland restoration and management in this water-limited region.