Hydrodynamics of nw iberian peninsula under past and future climate conditions

Resumo

The oceans are areas of high primary productivity which vary geographically. In general, coastal areas are more productive and host greater biodiversity than the open ocean. Coastal regions are usually limited to the continental shelf area and include areas of coastal upwelling. In these regions, ocean circulation is mainly influenced by topography and coastal winds. Coastal regions, being highly productive, concentrate a high percentage of the world's population and therefore have an essential socio-economic relevance. The Northwest coast of the Iberian Peninsula (NWIP) is one of the areas of the world with the highest primary productivity since several factors converge in this area: (1) it is located at the northern limit of the Canary Islands upwelling system, (2) a large number of large rivers flow into it, providing nutrients, and (3) it has a peculiar morphology with the presence of the Galician estuaries to the north. Over the years, numerous scientific investigations have been carried out in various fields to try to better understand the hydrodynamic behaviour of the NWIP coast. Despite this, many questions remain unanswered, and even more so if we take into account the current scenario of climate change, both natural and anthropogenic, that we are facing. The study of the effects of climate change on the NWIP coast is a difficult but necessary task, as it helps to understand its vulnerabilities and allows planning strategies for mitigation and adaptation. Numerical models are very useful tools for carrying out this task, since they allow us to simulate, for example, the water temperature in the future under various climate change scenarios. In addition, the models also allow us to simulate historical events in order to analyze them in detail. For all the above described, the main objective of this thesis is to contribute to the oceanographic knowledge of the NWIP coast and help to evaluate the impact that climate change may have on the region. For this purpose, the numerical model Delft3D was used, which allows improving the spatial resolution of global climate models that cover larger areas, in order to solve regional and even local processes in a more accurate way. The NWIP coast is a complex system in which the morphological characteristics of the coast, the bathymetry, the presence of freshwater from the discharge of the main rivers and the wind regime that controls the upwelling and downwelling cycles modifies the general circulation. The complexity of the system makes the application of numerical models, such as Delft3D, to this area an arduous task. First, it is necessary to generate a mesh whose resolution is adequate to reproduce the hydrodynamics of the area properly. Also, it must be taken into account that the simulations require a high computational cost, so it is necessary to reach a compromise between these factors. This thesis was carried out using two different meshes, a multi-domain one covering the NWIP from 10.20 to 8ºW and from 40 to 43.75ºN, and a single-domain one covering from 10.00ºW to 8.33ºW and from 41.18ºN to 43.50ºN. The multi-domain mesh is composed of five interconnected domains whose horizontal resolution varies, with the coastal domains having a higher resolution than the ocean domain. This configuration was used to assess the effect of climate change on the upwelling system. The single-domain mesh was used to carry out specific studies in the area of the Rías Baixas and the Minho River estuary. Once the mesh is created, it is necessary to calibrate the model. During the calibration process, the model is run, the results are compared with field data or results from other validated simulations, the calibration parameters are modified, the model is run again, the results are again compared with field data or results from other validated simulations and the process is repeated until the simulation results are sufficiently accurate. At this point, the model is calibrated and can be used in the study area. In the case of this thesis, the model has been calibrated by comparing the outputs with in situ data on sea surface elevation, salinity, temperature and horizontal speed near the surface. The capacity of the model to reproduce the dynamics of coastal systems, estuaries and rias, adjacent coast and the water exchange between the coastal systems and the adjacent coast, was evaluated by analyzing the intrusion of the Minho River plume into the Ría de Vigo. The Minho River flows approximately 30 km south of the Ría de Vigo, and previous studies have observed that, under conditions of southern winds, its plume is capable of reaching the Rías Baixas, entering them and modifying their typical circulation pattern. The Ría de Vigo is the southernmost estuary of the Rías Baixas and the one that is most influenced by the plume of the Minho River. The circulation pattern of the Rías Baixas and in particular of the Ría de Vigo, is a typical estuary circulation pattern, with fresh water coming out of the surface and salty ocean water entering through the deep layers. This pattern can be reversed under the influence of winds from the south and by a moderate to high discharge from the Minho River, as the winds push the river plume towards the coast, favouring it reaches the rias. To detect the intrusion of the plume of the Minho River into the Ría de Vigo, we looked for those days when the surface salinity at the southern mouth of the ria was lower than the values measured inside the ria. This pattern indicates freshwater intrusion from the platform into the ria. Weekly data from vertical salinity profiles were used for this purpose during the period 2006-2017. Field data showed that this pattern of intrusion was detected several consecutive weeks during the months of January and February 2010. Therefore, these dates were selected to perform the numerical simulations in order to determine if it was a single event or several consecutive pulses. In addition, we proceeded to characterize a complete intrusion event from the cross and longitudinal sections of density and horizontal velocity. This study made it possible to determine that the Minho River plume reaches the Ría de Vigo approximately 12 hours after the favourable wind peak, the intrusion events have a typical duration of 1.5 days and that they affect from the surface to ~10-15 m depth within the ria. Furthermore, during the intrusion event, the influence of the inner river, the Verdugo-Oitavén, on the circulation of the estuary is negligible. The numerical model was also validated for use when the boundary conditions come from climate models such as those implemented in the framework of the projects "Coupled Model Intercomparison" version 5 (CMPI5) and "Coordinated Regional Climate Downscaling Experiment" (CORDEX). These projects provide climate variables from numerous models executed by different institutions. As it was not feasible to run Delft3D from the variables obtained from all the models that comprise these projects, one was chosen. Statistical analysis was carried out to determine which CORDEX model best reproduces the variables used to force the Delft3D model. The comparison was made between the outputs of the climate models and the ERA-Interim reanalysis data for the historical period. Thus, it was determined that the model that best reproduces the variables analyzed for the study area is the MOHC-HadGE2-Es, and therefore, its outputs were used to force the Delft3D model. Once the model was validated, and the data sources to force the Delft3D model for future projections were chosen, the impact of climate change in the study area was analyzed. Future projections were performed under the RCP8.5 greenhouse gas emission scenario, which is currently considered the most likely. Using the multi-domain grid, how climate change may affect the upwelling was evaluated. The objective was to reproduce an upwelling event from intense to extreme characteristic of the historical and future periods. For this purpose, the upwelling index (UI) was calculated during July and August, a time of intense upwelling, in several points along the coast for the historical and future period. Once the UI was obtained, values between the 75% and 99% percentile were chosen as representative of events between intense and extreme and were used to force the model. The UI values obtained are higher for the future than for the historical period. During an upwelling event, cold nutrient-rich water rises to the surface, so the efficiency of an upwelling event can be measured in terms of its ability to reduce the surface layer temperature. Thus, the differences in surface temperature (SST) between the end and the beginning of the upwelling event were calculated (ΔSST= SSTend - SST beginning), and the results obtained for the historical and future events were compared. The more negative ΔSST, the higher capacity of the upwelling to pump deep water to the surface and therefore, the more effective it will be. The differences in the TSS for the historical and future events show similar results in most of the domain, except for the region between 41º 30' N and 42º 15' N, dominated by the main rivers (Minho, Douro and Lima). Near the coast, the drop in water temperature range from 2 to 3 ºC in the historical event and from1 to 2 ºC in the future. Although IU values are higher for the future period, so it could be expected that upwelling will be more intense, simulations suggest that upwelling events will be less effective in the future. The analysis of the stratification of the water column, carried out in terms of the Brunt-Väisälä frequency calculated for sections between the Ría de Vigo and the Minho estuary, indicates that it will increase throughout the section by the end of the century. The thermocline will be located at greater depth, and the gradient will be more marked. In this way, the increase in the stratification counteracts the intensification of the upwelling favourable winds. This could have negative effects on the primary production of the NWIP coast since a reduction of the upwelling effectiveness implies a reduction of the vertical mixing and therefore of a reduction in the transport of nutrients and oxygen through the water column. In addition, the decrease in SST due to the rise of cold water is acting as a buffer to ocean warming for many species, which find refuge in upwelling areas from ocean warming. At the beginning of this summary, it was indicated that the coastal areas are areas of high socio-economic importance, in the Rías Baixas, in particular, there are many companies linked to sectors such as fishing, aquaculture and shellfish farming. The mussel aquaculture sector is one of the most important in the world. In 2018, 279000 tons of mussels were produced, representing 40% of European production and more than 15% of world production. Mussel cultivation is carried out extensively, with the largest number of rafts being found in the Ría de Arousa. The cultivated species is Mytilus galloprovincialis, and it is grown on ropes that are usually 12 m long. The growth and mortality of mussels depend on many environmental factors, such as the availability of oxygen and phytoplankton, water temperature, salinity and water pH, among others. Water temperature is the most relevant factor as it can explain 67% of the differences in growth. Previous studies of Mytilus galloprovincialis determined that the optimal temperature range for its growth is between 14 and 20 ºC. In addition, research on the possibility of adaptation of mussels to the expected increase in water temperature indicates that, this optimal temperature range is not expected to vary significantly. Although they survive in a wide range of temperatures, they are close to their physiological limit of thermal tolerance. Therefore, the current optimal temperature range for mussel growth was used to determine the comfort index both in the historical period and for the end of the century. The comfort index was defined as the percentage of time during which the water temperature remains within the optimal temperature range for mussel growth. In this case, the model was run using the single domain grid for the months of July and August of 20 historical years (1999-2018) and future years (2080-2099). The water temperature values were used in the location of each of the mussel rafts polygons that are currently distributed in the Rías Baixas to calculate the comfort index. Since the ropes where the mussels are suspended measure a maximum of 12 m, the water column was divided into two layers, the surface layer from 0 to 6 m and a deep layer from 6 to 12 m, and the comfort index was calculated for each of them. The results show that for the historical period the comfort index is 100% in all the polygons, both in the surface layer and in the deep layer. This means that the mussels are under optimal temperature conditions 100% of the time, which favours their growth. For this reason, the Rías Baixas are an ideal place for mussel aquaculture as indicated by the large production they support today. By the end of the century, projections suggest that in the majority of the mussel farms the comfort index in the superficial layers will be reduced by more than 60% and around 30% in the deep layers. The outermost areas of the rias are the least affected, with points in the deep layer where comfort would still be at 100% by the end of the century. This is because these are areas currently close to the lower limit of the optimum temperature range (14 ºC) as the upwelling highly influences them. In these areas, even if the water temperature increases in the future, the rise in water temperature will not be sufficient to exceed the upper limit (20ºC). Stratification is another physical variable of great importance for mussel production since, as indicated above, the vertical exchange of nutrients and oxygen is limited in a highly stratified water column. The comparison of Brunt-Väisälä frequency at each of the trough polygon points for the historical period and future projections shows that by the end of the century stratification will increase in most of the polygons. This stratification will only be reduced in the innermost part of the rias. This is because these areas are highly influenced by the inner river, which favours a haline stratification. In the future, a reduction in river flow of 25% is predicted under a RCP 8.5 scenario, so introducing this consideration into the model results in a reduction in the stratification of haline origin. Finally, it was evaluated how climate change may affect two species of habitat-forming macroalgae (Himantalia elongata and Bifurcaria bifurcata) currently found in the Rías Baixas. Habitat-forming macroalgae are species of great ecological interest as they provide structure, shelter and food for many accompanying species, forming ecological communities. Currently, a reduction and contraction of the populations of these species is being observed due to the increase in the temperature of the sea. On the coasts of the Iberian Peninsula, a decline in the populations on the Cantabrian coast has been observed. However, areas such as the Rías Baixas are acting as contemporary climatic refuges for these species as they are protected areas from the action of the waves and points where the upwelling keeps the water colder than in the adjacent coastal areas while providing it with a large amount of nutrients. Given the current situation, it is necessary to analyze whether or not the estuaries will continue to act as a climatic refuge in the future. For this purpose, the results of surface water temperature for July and August of the historical period (1999-2018) and the future (2080-2099) were used. Previous studies determined that the thermal survival threshold for H. elongata is 18 ºC and for B. bifurcata 24.7 ºC sustained, in both cases, for at least 10 consecutive days. A mechanistic model of species distribution using the thermal survival threshold of H. elongata was developed by calibrating it with field data. This species was chosen because it shows less prevalence in the study area. This model is based on the percentage of time that the algae are able to survive under lethal conditions. Once the species distribution model was calibrated, it was used to perform thermal suitability maps for each of the macroalgae both at present and by the end of the century. Currently, thermal conditions are favourable for the presence of both macoalgae in most of the rias. A detailed analysis shows that the rias of Muros and Arousa have the most suitable conditions for the presence of H. elongata. However, in the central part of the Ría de Arousa they are less favourable when compared to the outer part. In the Rías de Pontevedra and Vigo the conditions are less favourable than in the two more northern rías. The records of presence and absence of H. elongata and other macroalgae with similar thresholds corroborate these results since fewer populations are observed in the two southern rías. Projections for the end of the century indicate that thermal conditions will be lethal for H. elongata, and probably for other macroalgae with similar thermal thresholds. However, the analysis of thermal suitability of the habitat for B. bifurcata, characterized by a higher thermal threshold of survival than H. elongata, shows that the Rías are thermally suitable for its presence and will generally continue to be so at the end of the century. Furthermore, due to the upwelling, the rias will continue to be areas where the water temperature will be lower than that of the adjacent ocean. Therefore they may continue to be considered thermal refuges for species such as B. bifurcata and those with a similar thermal threshold.