The remarkable salinity difference between the upper Pacific and Atlantic Oceans is often explained through net export of water vapour across Central America. To investigate this mechanism a study of salinity signals in the Equatorial Pacific Ocean current system was made looking at responses to fresh water input from two sources (local versus remote - Atlantic Ocean) as well as a combination of the two. Statistical analyses (Empirical Orthogonal Functions, Single Value Decomposition and Wavelet analysis) were used to split the main sources of the atmospheric freshwater input into local and remote contributions and to quantify both contributions. The remote source was assumed to have been transported over Central America from the Atlantic Ocean as an atmospheric freshwater flux, whereas the local source originated in the Pacific Ocean itself. The analysis suggests that 74% of the total variance in precipitation over the tropical eastern Pacific is due to water vapour transport from the Atlantic. It also demonstrates strong influence of ENSO events, with maximum correlation at a two months time lag. During La Ni�a periods the precipitation variance is more closely related to water vapour transport across Central America (the remote source), while during El Ni�o periods it is more closely related to the water vapour transport by Southerly winds along the west coast of South America (the local source). The current and temperature fields provided by the Modular Ocean Model (version 2) were used to study the changes in the salinity field when freshwater was added to or removed from the model. ECMWF ERA-40 data taken from the ECMWF data server was used to determine the atmospheric flux of freshwater at the ocean surface, in the form of evaporation minus precipitation (E-P). The Mixed Layer Depth (MLD) computed from temperature and salinity fields determines to what depth the salinity's dilution/concentration takes place for every grid point. Each MLD was calculated from the results of the previous time step, and the water column was considered well mixed from the surface to this depth. The statistical relationships were used to reconstruct the precipitation over the tropical eastern Pacific. A numerical ocean model, which uses currents and temperature from a global ocean model and is forced by precipitation, was used to study the ocean's response to either the remote or the local source acting in isolation. Through time lag correlation analysis of the sea surface salinity anomalies produced by the variation in the reconstructed precipitation fields, it is found that the anomaly signals of salinity propagate westward along the Equator at a rate of approximately 0.25 m.s-1 (6.1 degrees per month).

The European Space Agency's Soil Moisture and Ocean Salinity (SMOS) satellite was launched on November, 2, 2009 from the Russian cosmodrome of Plesetsk. Its objective is to globally and regularly collect measurements of soil moistre and Sea Surface Salinity (SSS). To do that, a pioneering instru- ment has been developed: the Microwave Imaging Radiometer by Aperture Synthesis (MIRAS), the rst space-borne, 2-D interferometric radiometer ever built; it operates at L-band, with a central frequency of 1.4135 GHz, and consists of 69 antennas arranged in a Y shape array. MIRAS' output are brightness temperature maps, from which SSS can be derived through an iterative algorithm, and using auxiliary information. For each overpass of the satellite an SSS map is produced, with an estimated accuracy of 1 psu (rmse). According to the Global Ocean Data Assimilation Experiment (GODAE) the mission requirement is instead speci ed as 0.1 psu after av- eraging in a 10-day and 2 2 spatio-temporal boxes. In previuos works ((Sabia et al., 2010), or more extensively in Dr. Sabia's Ph.D. thesis (Sabia, 2008)) the main error sources in retrieving SSS from SMOS measurements were determined as: 1. Scene-dependent bias in the simulated measurements, 2. L-band forward modeling de nition, 3. Radiometric sensitivity and accuracy, 4. Constraints in the cost function, and 5. Spatio-temporal averaging. This Ph.D. thesis, is an attempt of reducing part of the aforementioned errors (the relative to the one-overpass SSS (1 - 4)) by a more sophisticated data processing. Firstly, quasi-realistic brightness temperatures have been simulated using the SMOS End-to-end Performance Simulator (SEPS) in its full mode and an ocean model, as provider for geophysical parameters. Using this data set the External Brightness Temperature Calibration technique has been tested to mitigate the scene-dependent bias, while the error introduced by inaccuracies in the L-band forward models has been accounted for by the application of the External Sea Surface Salinity Calibration. Apart from simulated brightness temperatures, both External Brightness Temperature Calibration and External Sea Surface Salinity Calibration have been tested using real synthetic-aperture brightness temperatures, collected by the Helsinki University of Technology HUT-2D radiometer during the SMOS Calibration and Validation Rehearsal Campaign in August 2007 and ten days of data acquired by the SMOS satellite between July 10 and 19, 2010. Finally, a study of the cost function used to derive SSS has been performed: the correlation between measurement mis ts has been estimated and the e ect of including it in the processing have been assessed. As an outcome of a 3-month internship at the Laboratoire LOCEAN in Paris, France, a theoretical review of the e ect of the rain on the very top SSS vertical pro le has been carried out and is presented as Appendix.

An analysis to determine physical and spatial patterns of the surface latent heat flux (LHF) and near surface (5m) salinity (NSS) beneath tropical cyclones (TCs) in the North Atlantic and eastern North Pacific basins during the first 24 hours of rapid intensification (RI) was conducted using empirical orthogonal function (EOF) analysis. To determine if these patterns were unique to RI, TC RI cases were compared to three non-RI intensification thresholds, 10 kt, 15 kt and 20 kt, for both LHF and NSS. Though similarities exist between non-RI and RI cases physical and spatial patterns unique to the RI cases did exist. Sea surface temperatures associated with statistically identified TC groups were assessed for their potential influence on RI. While inconclusive in the eastern North Pacific, NSS in the Atlantic may play a role for RI TCs in areas affected by river discharge from South America.

A monthly resolved coral δ18O record from Sabine Bank, Vanuatu (SBV; 166.04° E, 15.94°S), extending from 2006 to 1929 CE, is used to assess the influence of sea surface salinity (SSS) on the oxygen isotopic composition of coral aragonite at this location. Monthly SSS anomalies at SBV between 2006 and 1970 are strongly correlated with monthly anomalies in sea surface temperature (SST) variations in the central Pacific cold tongue, as recorded by SST anomalies in the Niño 3.4 grid box (i.e., canonical record of ENSO variability, r = 0.68, p < 0.01; lag of 6 months). This relationship demonstrates that SSS in the waters offshore of Vanuatu respond to ENSO-driven changes in the coupled ocean-atmosphere system in the tropical Pacific. SBV coral δ18O is also strongly correlated with monthly instrumental SSS anomalies at Vanuatu (r = 0.71, p < 0.01), therefore SBV coral δ18O variations are driven by the ENSO-related changes in surface ocean conditions. A calibration-verification exercise using SBV coral δ18O values and instrumental SSS was performed over the period 2006-1970 CE. A statistically robust transfer function was determined and used to predict SSS at SBV back to 1929 CE. The coral δ18O and SSS relationship at Vanuatu is further evaluated via comparison with a coral δ18O record from Malo Channel, Vanuatu, a site that is 130 km to the east of SBV. The strong correlation between the two coral δ18O records (r = 0.70; p < 0.01) suggests that ENSO drives regional changes in SSS in this region and that such changes can be reconstructed using variations in skeletal δ18O of corals. / text

This thesis contributes to the development of dinoflagellate cysts as indicators of past environmental change in the Northeastern Pacific coastal ocean, and investigates past variations in sea-surface temperature, salinity and primary productivity encoded in dinoflagellate cyst sedimentary records from the Santa Barbara Basin (SBB, southern California) and Effingham Inlet (Vancouver Island, British Columbia) over the last millennium. The dinoflagellate cyst records extracted from the SBB and Effingham Inlet predominantly laminated sediments and analysed at sub-decadal resolutions, constitute some of the most detailed records of cyst-producing dinoflagellate populations in the world. A two year-long sediment trap study from the SBB documents the seasonality in dinoflagellate cyst production for the first time on the Pacific coast of the United States. The study shows that dinoflagellate cyst data can be used as indicators of changes in sea-surface temperature and primary productivity associated with seasonal upwelling in the SBB. In particular, several dinoflagellate cyst taxa such as Brigantedinium spp. and Lingulodinium machaerophorum are identified as indicators of “active upwelling” (typically occurring in spring and early summer) and “relaxed upwelling” conditions (fall and early winter) at the site, respectively. Analysis of a dinoflagellate cyst record from the SBB spanning the last ~260 years at biannual resolution documents the response of cyst-producing dinoflagellates to instrumentally-measured warming during the 20th century, and reveals decadal scale variations in primary productivity at the site that are coherent with phases of the Pacific Decadal Oscillation (PDO). The cyst assemblages are dominated by cysts produced by heterotrophic dinoflagellates (in particular Brigantedinium spp.), but the turn of the 20th century is marked by an abrupt increase in concentrations of L. machaerophorum and Spiniferites ramosus, two cyst taxa of autotrophic affinity. Their increasing abundances during the 20th century are interpreted to reflect warmer conditions and possibly stronger stratification during summer and fall. The dinoflagellate cyst data suggest a warming pulse in the early 1900s and provide further evidence that persistently warmer and/or more stratified conditions were established by the late 1920s. The dinoflagellate cyst record from Effingham Inlet, spanning the last millennium, is characterized by the proportionally equal contribution of cysts produced by autotrophic and heterotrophic dinoflagellates in most samples. The cyst data indicate variations in sea-surface temperature, salinity and primary productivity that are associated with local expressions of the Medieval Climate Anomaly (from the base of the record to ~1230), the Little Ice Age (from ~1230 to ~1900) and warming during the second half of the 20th century. Both dinoflagellate cyst records reveal that since the beginning (in the SBB) and mid-20th century (in Effingham Inlet), autotrophic dinoflagellates contribute to a greater portion of the primary production in the region, whereas heterotrophic dinoflagellates, as indicators of diatom populations, decline. Variability in the dinoflagellate cyst data is coherent at both sites and suggest a reduced expression of decadal scale variability associated with the PDO during the 19th century. / Graduate / 0416 / 0427 / mbringue@uvic.ca

Previous studies showed that sea surface salinity (SSS) in the Florida Straits as well as Florida Current transport covaried with changes in North Atlantic climate over the past two millennia. However, little is known about earlier Holocene variability in the Florida Straits. Here, we combine Mg/Ca-paleothermometry and stable oxygen isotope measurements on the planktonic foraminifera Globigerinoides ruber (white variety) from Florida Straits sediment core KNR166-2 JPC 51 (24 degrees 24.70? N, 83 degrees 13.14?W, 198m deep) to reconstruct a high-resolution (~30 yr/sample) early to mid Holocene record of sea surface temperature and delta18OSW (a proxy for SSS) variability. We also measured Ba/Ca ratios in the same shell material as a proxy for riverine input into the Gulf of Mexico over the same time interval. After removing the influence of global delta18OSW change due to continental ice volume variability, we propose that early Holocene SSS enrichments were caused by increased evaporation/precipitation ratios in the Florida Straits associated with periods of reduced solar output, increased ice rafted debris in the North Atlantic and the development of more permanent El Nino-like conditions in the eastern equatorial Pacific. When considered with previous high-resolution reconstructions of early Holocene tropical atmospheric circulation changes, our results provide evidence that solar output variability over the Holocene had a significant impact on the global tropical hydrologic cycle over the last 10,000 years.

Deep ventilation of the Black Sea is inhibited by a sharp salinity gradient within the upper water column, resulting in a shallow anoxic interface at around 100 &ndash / 200 m depth. Understanding biological and chemical processes within the boundary region between oxic and anoxic waters is fundamental to comprehend the biogeochemical response of the Black Sea to climate forcing. The structure and depth of the chemocline is largely determined by the physical processes which transport surface waters to depth. Here we investigate how the structure and stability of the upper water column responds to changes in climatic forcing over interannual to multidecadal time-scales. We report results from two hydrodynamic model reanalysis. The first, extending from 1971-1993 assimilates CTD data. The second, extending from 1992-2001, assimilates altimetry data. Model results are validated against CTD and satellite data and consistency between modeled surface properties and observations is demonstrated. A problem with the data assimilation scheme of the 1992 -2001 model run is identified, which results in model drift and an unrealistic water column structure at intermediate depths. Model results indicate a warming trend of 0.7 &deg / C in sea surface temperature and a freshening trend of 0.4 in sea surface salinity between 1971 and 2001, with an associated increasing trend in the stability of the seasonal thermocline and a declining trend in surface mixed layer depth of 6.3 m. Trends are superimposed on a distinct multiannual variability characterized by relatively warm and saline conditions between 1971 and 1984, relatively cool and fresh conditions between 1985 and 1993 and warm and fresh conditions post-1993. The period between 1985 and 1993 corresponds to higher NAO and EA/WR index values although these indices do not exhibit a similar ~decadal scale variability. Higher frequency interannual variability in water column characteristics is related to the NAO and EA/WR atmospheric indices. Despite the cool conditions prevalent during the 1990s, the persistent freshening trend caused a reduction in the density of mixed layer waters throughout the study period. A positive feedback is proposed between increasing SSTs, reduced vertical mixing and freshening of the surface layer which further increases the stability of the upper water column. CIL characteristics typically mirrored surface temperature characteristics and varied considerably between the relatively warm period during the early part of the study and the subsequent cool period. The mean thickness and temperature of the CIL between 1971 and 1981 were ~39 m and ~7.5 &deg / C respectively, as compared to ~47 m and ~7.4 &deg / C between 1982 and 1993. Freshening of the upper water column also resulted in an increase in the stability maxima that exists at the base of the CIL, suggesting reduced ventilation of the upper water column during winter.

Dans cette thèse, nous avons abordé plusieurs thématiques liées aux modes de variabilité climatique dans l'Atlantique tropical à l'échelle interannuelle. Les analyses statistiques nous ont permis dans un premier temps de mettre en évidence les deux principaux modes dominants de cette variabilité interannuelle: un mode équatorial et un mode méridien. Le mode équatorial est responsable d'anomalies de température de surface de la mer (SST) principalement dans le Golfe de Guinée et est identifié par des variations de la pente du niveau de la mer dans la bande équatoriale. Il est dû à des rétroactions dynamiques entre le vent, le niveau de la mer et la SST. Quant au mode méridien, il se manifeste par des fluctuations inter-hémisphériques de SST et est contrôlé par des rétroactions dynamiques et thermodynamiques entre le vent, l'évaporation et la SST. L'évaluation du couplage de ces variables clés du mode méridien nous a permis de proposer un modèle conceptuel pour expliquer les principaux mécanismes responsables des oscillations du mode méridien. Le modèle a montré que le mode méridien résulte de la superposition d'un mécanisme auto-entretenu basé sur les rétroactions positives et négatives générant des oscillations régulières de haute fréquence (2-3 ans) et d'un autre mécanisme d'oscillation basse fréquence (4-9 ans) lié à l'influence d'ENSO du Pacifique Est. Comme l'évolution de ces deux modes est fortement liée au déplacement méridien de la zone de convergence intertropicale (ITCZ) qui transporte les pluies, nous avons ensuite identifié la signature de ces modes sur la salinité de la surface de la mer à l'aide observations in situ et d'une simulation numérique régionale. Les processus océaniques et/ou atmosphériques responsables de la signature de chaque mode ont été également identifiés grâce à un bilan de sel dans la couche de mélange du modèle validé. Le bilan de sel a révélé que le forçage atmosphérique, lié à la migration de l'ITCZ, contrôle la région équatoriale tandis que l'advection, due à la modulation des courants, du gradient vertical et le mélange à la base de la couche de mélange, explique les variations de SSS dans les régions sous l'influence des panaches. [...] / In this thesis, we investigate several topics related to the interannual climatic modes in the tropical Atlantic. Statistical analyses allows us to extract the two main dominant modes of interannual variability: an equatorial mode and a meridional mode. The equatorial mode is responsible for Sea Surface Temperature (SST) anomalies mainly found in the Gulf of Guinea and is linked to variations of the sea-level slope in the equatorial band. It is due to dynamic feedbacks between zonal wind, sea level and SST. The meridional mode is characterised by inter-hemispheric SST fluctuations and is controlled by dynamic and thermodynamic feedbacks between the wind, evaporation and SST. After quantifying the coupling between key variables involved in the meridional mode, we develop a conceptual model to explain the main mechanisms responsible for meridional mode oscillations. The model shows that the meridional mode results from the superposition of a self-sustaining mechanism based on positive and negative feedbacks generating regular oscillations of high frequency (2-3 years) and another low frequency oscillation mechanism (4-9 years) related to the influence of ENSO. As the evolution of these two modes is strongly linked to the meridional shift of the Intertropical Convergence Zone (ITCZ) and associated rainfall maximum, we identify the signature of these modes on Sea Surface Salinity (SSS) using in situ observations and a regional numerical simulation. Oceanic and/or atmospheric processes responsible for the signature of each mode are also identified through a mixed-layer salt budget in the validated model. The salt balance reveals that the atmospheric forcing, related to the ITCZ migration, controls the equatorial region while the advection, due to the modulation of current dynamics, the vertical gradient and mixing at the base of the mixed layer, explains SSS variations in regions under the influence of plumes. Finally, we study the Equatorial Kelvin wave characteristics and influences on the density that are involved in the meridional and equatorial mode connection processes, using a very simplified model of gravity wave propagation along the equator. After a brief description of this model, which was initially constructed to study dynamics in the equatorial Pacific, we apply it to the specific case of the equatorial Atlantic by validating its analytical and numerical solutions under adiabatic conditions. [...]