One major acknowledged challenge in daily precipitation is the inability to model extreme events in the spectrum of events. These extreme events are rare but may cause large losses. How to realistically simulate extreme behavior of daily precipitation is necessary and important. To that end, a hybrid probability distribution is developed. The logic of this distribution is to simulate the low to moderate values by an exponential distribution and extremes by a generalized Pareto distribution. Compared with alternatives, the developed hybrid distribution is capable of simulating the entire range of precipitation amount and is much easier to use. The hybrid distribution is then used to construct a bivariate discrete-continuous mixed distribution, which is used for building a daily precipitation generator. The developed generator can successfully reproduce extreme events. Compared with other widely used generators, the most important advantage of the developed generator is that it is apt at extrapolating values significantly beyond the upper range of observed data. The major challenge in monthly streamflow simulation is referred to the underrepresentation of inter-annual variability. The inter-annual variability is often related with sustained droughts or periods of high flows. Preserving inter-annual variability is thus of particular importance for the long-term management of water resources systems. To that end, variables conveying such inter-annual signals should be used as covariates. This requires models that must be flexible at incorporating as many covariates as necessary. Keeping this point in mind, a joint conditional density estimation network is developed. Therein, the joint distribution of streamflows of two adjacent months is assumed to follow a specific parametric family. Parameters of the distribution are estimated by an artificial neural network. Due to the seasonal concentration of precipitation or the joint effect of rainfall and snowmelt, monthly streamflow distribution sometimes may exhibit a bimodal shape. To reproduce bimodality, nonparametric models are often preferred. However, the simulated sequences from existing nonparametric models represent too close a resemblance to historical record. To address this issue, while retaining typical merits of nonparametric models, a multi-model regression-sampling algorithm with a few weak assumptions is developed. Collecting hydrometric data is the first step for building hydrologic models, and for planning, design, operation, and management of water resource systems. In this dissertation, an entropy-theory-based criterion, termed maximum information minimum redundancy, is proposed for hydrometric monitoring network evaluation and design. Compared with existing similar approaches, the criterion is apt at finding stations with high information content, and locating independent stations.

主催：JST/CREST，Vrije University, ALTERRA, IBPC

Coast redwood (Sequoia sempervirens) ecosystems are strongly influenced by the presence of summer marine fog, and variation in fog frequency is closely linked to climate variation in the NE Pacific region. Because oxygen isotope composition (𝛿¹⁸O) of organic matter records distinct water sources (e.g. summertime fog vs. winter precipitation) and carbon isotopes (𝛿¹³C) are typically sensitive to humidity and water status, it then follows that inter-annual variation in tree-ring isotope ratios, which are coherent across multiple sites, should preserve a potentially powerful proxy for climate reconstruction. Here we present an analysis of a 50-year time series for both 𝛿¹⁸O and 𝛿¹³C values from subdivided tree rings obtained from multiple redwood trees at multiple sites. Within-site and between site correlations were highly significant (p < 0.01) for the 𝛿¹⁸O time series indicating a regionally coherent common forcing of 𝛿¹⁸O fractionation. Within-site and between-site correlation coefficients were lower for the 𝛿¹³C than for the 𝛿¹⁸O time series although most were still significant (at least to p < 0.05). The hypothesized reason for the differences in the correlation is that carbon isotope discrimination is more sensitive to microenvironmental and tree-level physiological variation than is 𝛿¹⁸O fractionation. Stable-isotope variation in tree-ring cellulose was similar between slope, gully and riparian micro-habitats within a single watershed, implying that minor topographic variation when sampling should not be a major concern. These results indicate that stable-isotope time series from redwood tree rings are strongly influenced by regional climate drivers and potentially valuable proxies for Pacific coastal climate variability.

The modeling results of current global aerosol models agree, generally within a factor of two, with the measured surface concentrations of black carbon (BC) and sulphate (SF) aerosols in rural areas across the northern continents. However, few models are able to capture the observed seasonal cycle of the Arctic aerosols. In general, the observed seasonality of the Arctic aerosols is determined by complex processes, including transport, emissions and removal processes. In this work, the representations of aerosol deposition processes (i.e., dry deposition, in-cloud and below-cloud scavenging) within the framework of the Canadian Global Air Quality Model – GEM-AQ are first enhanced. Through the enhancements in GEM-AQ, the seasonality of the Arctic BC and SF is reproduced, and the improvement in model performance extends to the rest of the globe as well. Then, the importance of these deposition processes in governing the Arctic BC and SF seasonality is investigated. It is found that the observed seasonality of the Arctic BC and SF is mainly caused by the seasonal changes in aerosol wet scavenging, as well as the seasonal injection of aerosols from surrounding source regions. Being able to reproduce the seasonality of the Arctic BC, the enhanced GEM-AQ allows more accurate assessment of the contributions of anthropogenic sources to the BC abundance in the Arctic air and deposition to the Arctic surface. Simulating results on regional contributions to the Arctic BC show a strong dependence on altitude. The results reinforce the previous finding of Eurasia being the dominant contributor to the surface BC in the Arctic, and suggest a significant contribution from Asian Russia. In addition to the seasonality of the Arctic aerosols, the inter-annual variation in the Arctic BC surface concentration is also investigated. To complement the 3-D GEM-AQ model, the atmospheric backward trajectory analysis, together with estimated BC emissions, is implemented as a computational effective approach to reconstruct BC surface concentrations observed at the Canadian high Arctic station, Alert. Strong correlations are found between the reconstructed and the measured BC in the cold season at Alert between 1990 and 2005, which implies that atmospheric transport and emissions are the major contributors to the observed inter-annual variations and trends in BC. The regional contributions estimated annually from 1990 through 2005 suggest that Eurasia is the major contributor in winter and spring to the near-surface BC level at Alert with a 16-year average contribution of over 85% (specifically 94% in winter and 70% in spring). A decreasing trend in the Eurasian contribution to the Arctic is found in this study, which is mainly due to regional emission reduction. However, the inter-annual variation in the North American contribution shows no clear trend.

Clear-sky top-of-atmosphere (TOA) fluxes are important in estimating the impact of clouds on our climate. In this study, we quantitatively compare the clear-sky fluxes measurements of the Clouds and the Earths Radiant Energy System (CERES) instrument to clear-sky fluxes from two reanalysis, NASA's Modern Era Retrospective-analysis for Research and Application (MERRA), and the European Center for Medium Range Weather Forecast Interim reanalysis (ERA-Interim). In the first comparison, we compare observed fluxes from individual cloud-free field-of- views to the reanalyses. In the second comparison, we compare monthly averaged observed clear-sky fluxes to those from the reanalyses. Monthly clear-sky fluxes are calculated by averaging fluxes from cloud-free regions. In both comparisons, the fluxes generally agree within +/- 10 W/m^2. Finally, we show that, while the differences between the fluxes of observations and the reanalyses are several W/m2, the inter-annual anomalies agree much better, with zonal and global average inter-annual anomalies typically agreeing within 1 W/m^2. The longwave clear-sky anomalies show excellent agreement even when comparing individual grid points, whereas the shortwave clear-sky anomalies are generally smaller at individual grid points.

The modeling results of current global aerosol models agree, generally within a factor of two, with the measured surface concentrations of black carbon (BC) and sulphate (SF) aerosols in rural areas across the northern continents. However, few models are able to capture the observed seasonal cycle of the Arctic aerosols. In general, the observed seasonality of the Arctic aerosols is determined by complex processes, including transport, emissions and removal processes. In this work, the representations of aerosol deposition processes (i.e., dry deposition, in-cloud and below-cloud scavenging) within the framework of the Canadian Global Air Quality Model – GEM-AQ are first enhanced. Through the enhancements in GEM-AQ, the seasonality of the Arctic BC and SF is reproduced, and the improvement in model performance extends to the rest of the globe as well. Then, the importance of these deposition processes in governing the Arctic BC and SF seasonality is investigated. It is found that the observed seasonality of the Arctic BC and SF is mainly caused by the seasonal changes in aerosol wet scavenging, as well as the seasonal injection of aerosols from surrounding source regions. Being able to reproduce the seasonality of the Arctic BC, the enhanced GEM-AQ allows more accurate assessment of the contributions of anthropogenic sources to the BC abundance in the Arctic air and deposition to the Arctic surface. Simulating results on regional contributions to the Arctic BC show a strong dependence on altitude. The results reinforce the previous finding of Eurasia being the dominant contributor to the surface BC in the Arctic, and suggest a significant contribution from Asian Russia. In addition to the seasonality of the Arctic aerosols, the inter-annual variation in the Arctic BC surface concentration is also investigated. To complement the 3-D GEM-AQ model, the atmospheric backward trajectory analysis, together with estimated BC emissions, is implemented as a computational effective approach to reconstruct BC surface concentrations observed at the Canadian high Arctic station, Alert. Strong correlations are found between the reconstructed and the measured BC in the cold season at Alert between 1990 and 2005, which implies that atmospheric transport and emissions are the major contributors to the observed inter-annual variations and trends in BC. The regional contributions estimated annually from 1990 through 2005 suggest that Eurasia is the major contributor in winter and spring to the near-surface BC level at Alert with a 16-year average contribution of over 85% (specifically 94% in winter and 70% in spring). A decreasing trend in the Eurasian contribution to the Arctic is found in this study, which is mainly due to regional emission reduction. However, the inter-annual variation in the North American contribution shows no clear trend.

The ongoing climate changes are predicted to accelerate fast in arctic regions with increases in both temperatures and precipitation. Although the duration of snow cover is generally expected to decrease in the future, snow depth may paradoxically increase in those areas where a large amount of the elevated precipitation will fall as snow. The annual distribution and duration of snow are important features in arctic ecosystems, influencing plant traits and species interactions in various ways. In this thesis, I investigated the effect of snow on plants and their interactions with herbivores by experimentally increasing the snow cover by snow fences in three different habitats along an environmental gradient in Abisko, northern Sweden. I found that the snow cover mattered for plant quality as food for herbivores and herbivore performance. An enhanced and prolonged snow cover increased the level of insect herbivory on dwarf birch leaves under field conditions. Autumnal moth larvae feeding on leaves that had experienced increased snow-lie grew faster and pupated earlier than larvae fed with leaves from control plots. These findings indicated that plants from snow-rich plots produced higher-quality food for herbivores. My studies showed that differences in snow-lie explained parts of the within-year spatial and seasonal variation in plant chemistry and patterns of herbivory in this arctic landscape. The relationship between leaf nitrogen concentration and plant phenology was consistent between treatments and habitats, indicating that snow per se, via a delayed phenology, was controlling the nitrogen concentration. The relationship between leaf age and level of herbivory was positive in the beginning of the growing season, but negative in the end of the growing season, indicating an increasing importance of plant palatability and a decreasing importance of exposure time in determining the level of herbivory throughout the growing season. The concentrations of phenolics varied among habitats, treatments and sampling occasions, suggesting that these plants were able to retain a mosaic of secondary chemical quality despite altered snow conditions. Furthermore, the nutrient limiting plant growth, according to N:P ratio thresholds, appeared to shift from nitrogen to phosphorus along the topographic gradient from snow-poor ridges to more snow-rich heathlands and fens. Snow addition had, however, no significant effect on other nutrient concentrations than nitrogen and no significant effect on the leaf N:P ratio, indicating that differences in snow cover could not explain the variation in plant nutrient concentrations among habitats. In a five-year study, I found opposing inter-annual effects of increased snow on plant chemistry. In contrast to earlier results, the effect of snow-lie on plant nitrogen concentration was predominantly negative. However, the effect of increased snow cover on the level of herbivory remained predominantly positive. The strong within-year relationship between snow-melt date (via plant phenology) and plant nitrogen concentration and level of herbivory could not predict inter-annual variation in the effect of snow manipulation. I did not find any conclusive evidence for a single factor causing the inter-annual opposing effect of snow addition, but the results indicated that interactions with summer and winter temperatures might be important. In conclusion, this thesis showed that climate-induced changes in snow conditions will have strong effects on plant traits and plant-herbivore interactions. However, alterations in snow cover do not influence all plant traits and the effect may vary in time and space.

The eddy-covariance method was used to measure net ecosystem productivity over three adjacent fields from 2009 to 2011: two annual cropping systems (oat-canola-oat and hay-oat-fallow) recently converted from perennial cropping, and a perennial hay/pasture. We compared the management practises, determined the net carbon budget, and examined the effects of inter-annual variability. Carbon accumulation began earlier in the spring and continued later in the fall at the perennial site, compared with the annual crop sites, due to a longer growing season and continual plant cover. Cumulative cropping season net ecosystem productivity at the perennial site ranged from 40 to 240 g C m^(-2) because of variable weather. Including harvest removals and manure additions, the perennial site gained 120 g carbon m^(-2) and the annual sites lost 240 and 415 g carbon m^(-2), respectively, over the 30-month period. This indicates that the annual cropping systems would decrease soil carbon at this location.

The eddy-covariance method was used to measure net ecosystem productivity over three adjacent fields from 2009 to 2011: two annual cropping systems (oat-canola-oat and hay-oat-fallow) recently converted from perennial cropping, and a perennial hay/pasture. We compared the management practises, determined the net carbon budget, and examined the effects of inter-annual variability. Carbon accumulation began earlier in the spring and continued later in the fall at the perennial site, compared with the annual crop sites, due to a longer growing season and continual plant cover. Cumulative cropping season net ecosystem productivity at the perennial site ranged from 40 to 240 g C m^(-2) because of variable weather. Including harvest removals and manure additions, the perennial site gained 120 g carbon m^(-2) and the annual sites lost 240 and 415 g carbon m^(-2), respectively, over the 30-month period. This indicates that the annual cropping systems would decrease soil carbon at this location.

Une étude inter-annuelle (2000-2003) de la distribution de l’abondance, la biomasse et la composition des espèces phyto-planctoniques couplée aux facteurs environnementaux et le méta zooplancton a été réalisée dans cinq bassins de la saline de Sfax avec un gradient de salinité croissant. Les résultats indiquent une similarité inter-annuelle concernant les caractéristiques physiques de l'eau à savoir la température et la salinité. Les concentrations des éléments nutritifs diminuent avec l'augmentation de la salinité. Nous avons enregistré une nette variation spatiale de la communauté phytoplanctonique. Ainsi, les diatomées dominent dans les premiers bassins A1, A16 et C2-1, suivies par les dino flagellés. Les chlorophycées dominent dans les bassins hypersalésM2 et TS. Les cyanobactéries ont été relativement abondantes dans les bassins M2 et TS. Les abondances et les biomasses phytoplanctoniques les plus élevées ont été enregistrées dans les bassins hyper salés dues à la prolifération de la chlorophycée Dunaliella salina. En raison de la stabilité des facteurs environnementaux,l'étude inter-annuelle de la succession du phytoplancton dans la saline de Sfax a montré des différences légères.Compte tenu du stress salin important, la communauté phytoplanctonique demeure au stade pionnier juvénile.Afin de mieux comprendre les interactions entre les paramètres abiotiques et biotiques, plusieurs analyses statistiques (Test de corrélation Pearson, ANOVA, ACP, Co- inerties….) ont été réalisées. Les relations entre les différents paramètres physiques, chimiques et biologiques au niveau du premier bassin à tendance marine A1 (de salinité 45 p.s.u) semblent être utile pour une modélisation écologique développée pour la saline de Sfax. Ce modèle mis au point a été développé en se basant sur la biomasse de la communauté phytoplanctonique en interaction avec les éléments nutritifs particulièrement l’azote et le phosphore. Une comparaison entre les résultats des simulations et les observations a montré que le modèle écologique a pure produire d’une manière satisfaisante la dynamique saisonnière et inter-annuelle du phytoplancton. Ainsi, les résultats du modèle écologique sont en concordance avec les observations. D’après les simulations, la diminution des concentrations des nutriments, ainsi que l’augmentation du débit d’entrée d’eau dans le bassin A1, entraînent une diminution de la biomasse phytoplanctonique. Cependant, et suite à un réchauffement climatique, une perturbation de la dynamique de la biomasse phytoplanctonique pourrait être observée. De plus, les simulations montrent qu’une année de faible pluviométrie accélère les efflorescences du phytoplancton. Le modèle établi dans la présente étude donne de nouvelles perspectives pour simuler des combinaisons entre les scénarii qui peuvent être dépendants. / The distribution of phytoplankton abundance, biomass and species composition coupled withenvironmental factors and metazooplankton was studied relatively intensively and over a period of fourconsecutive years in five ponds featuring a gradient of increasing salinity from near to that of sea water to a ninefoldconcentration from 2000 to 2003. The results indicate that the physical characteristics of the water(temperature and salinity) were quite similar over the years. The concentrations of nutrients decreasewith increasing salinity. The diversity of the phytoplankton community revealed a spatial variation. Thusdiatoms dominated in the first ponds A1, A16 and C2-1, followed by dinoflagellates. The Chlorophyceaedominated the hypersaline ponds M2 and TS. Cyanobacteria were relatively abundant in ponds M2 and TS. Thehighest phytoplankton abundances and biomasses were recorded in the hypersaline ponds, resulting from theproliferation of the Chlorophyceae Dunaliella salina. As a result of the stability of the environmental factors, theinter-annual study of the phytoplankton succession in the solar saltern of Sfax, revealed slight differences. Giventhe importance of salt stress the phytoplankton community remains in juvenile stage.To better understand the interaction between abiotic and biotic parameters, many statistical analyses(Pearson correlation test, ANOVA, PCA, Co-intertia….) were performed. The determination of the relationshipbetween physical, chemical and biological parameters in the first marine pond A1 (salinity: 45 p.s.u) appears tobe useful for the development of an ecological model of the solar saltern of Sfax. This model has been developedon the basis of the biomass of the phytoplankton community in relation with nitrogen and phosphorous nutrients.A comparison between simulation results and observations indicated that the ecological model was able toproduce a satisfactory seasonal and inter-annual phytoplankton dynamics. Thus, the results of the ecologicalmodel were in accordance with the observations. According to the simulations, the decrease of nutrientconcentrations and the increase of the water inflow into the pond A1, cause a decrease in phytoplankton biomass.Due to climate change, the disturbance of the dynamics of phytoplankton biomass was observed. Especially, thesimulations predict acceleration of phytoplankton blooms due to dry year. Finally, the model developed in thisstudy gives new insights in order to simulate a combination of independent scenarii.