Acceleration with multi-threading or graphics processing units is another possible direction for future research. For the second topic, there are many directions to improve the models or expand the results. Uncertainty in wind farm optimization is a major issue to address. Using uncertainty quantification to address the variability in the wind speed, wind direction, and performances of individual turbines will increase the confidence in optimal solutions. In addition, the inclusion of more models to represent the distribution of wind speeds at the rotor, tower and blade structures, rotor-nacelle assemblies, and cost models would greatly increase the fidelity of the models. More accurate wake models than the ones presented in this thesis are also under continued research, although may not be well suited for gradient-based optimization. To expand the results, more optimization studies where design variables are considered simultaneously should be considered. The computational cost of these models should also be addressed with parallelization, such as in the AEP calculation.Aquaculture is an important contributor to the Irish economy, producing products to the value of e167 million in 2016, including e105 million from farmed Atlantic salmon . Most Irish salmon farming is certified organic . Salmon farming in Ireland is associated with an intricate network of fish movements within and between the different types of salmon farms. There are three different farm types, including broodstock, freshwater, and seawater farms. In earlier work , horticulture solutions social network analysis was used in combination with spatial epidemiological methods to characterize the network structure of live farmed salmonid movements in Ireland.
It was demonstrated that characteristics of the network of live salmonid fish movements in Ireland would facilitate infection spread processes. These included a power-law degree distribution [that is, “scale free”], short average path length and high clustering coefficients [that is, “small world”], with the presence of farms that could potentially act as super-spreaders or super-receivers of infection, with few intermediaries of fish movement between farms, where infectious agents could easily spread, provided no effective barriers are placed within these farms . A small proportion of sites play a central role in the trade of live fish in the country. Similarly, we demonstrated that highly central farms are more likely to have a number of different diseases affecting the farm during a year, diminishing the effectiveness of in-farm bio-security measures , and that this effect might be explained by an increased chance of new pathogens entering into the farm environment . This is a very important area of research in aquaculture, especially considering that the spread of infection via fish movement is considered one of the main routes of transmission . Mathematical models and computer simulations offer the potential to study the spread of infectious diseases and to critically evaluate different intervention strategies . Through access to real fish movement data, these models can be programmed to incorporate both the time-varying contact network and data-driven population demographics. However, there are considerable computational challenges when stochastic simulations are conducted using livestock data, both computationally, including the need for efficient algorithms, and also with model selection and parameter inference .
An efficient modeling framework for event-based epidemiological simulations of infectious diseases has recently been developed , including the use of a framework that integrates within farm infection dynamics as continuous-time Markov chains and livestock data as scheduled events. This approach was recently used to model the spread of Verotoxigenic Escherichia coli O157:H7 in Swedish cattle . Cardiomyopathy syndrome is a severe cardiac disease of Atlantic salmon. It was first reported in the mid-1980s in farmed salmon in Norway and later detected in several other European countries, including the Faroe Islands , Scotland and, in 2012, in Ireland . CMS generally presents as a chronic disease, leading to long-lasting, low-level mortality, although some individuals experience sudden death. At times, however, CMS can present as an acute, dramatic increase in mortality associated with stress . A recent Norwegian study has identified risk factors for developing clinical CMS, including stocking time, time at sea, a previous outbreak of pancreatic disease or Heart and Skeletal Muscle Inflammation , and hatchery of origin . The economic impact of CMS is particularly serious as it occurs late in the life cycle, primarily during the second year at sea, by which time the incurred expenditure is high. No effective preventive measures are known, and there is no treatment available . In 2009, CMS was identified as a transmissible disease , and has been linked, in 2010 and 2011, to a virus resembling viruses of the Totiviridae family . The discovery of this virus, piscine myocarditis virus , has contributed to increased knowledge about the disease including the development of new diagnostic, research and monitoring tools . The agent is spread horizontally, between farms at sea, although there is some indication of a possible vertical transmission pathway .
Recent Norwegian research has shown that PMCV is relatively widespread, including in geographic regions and fish groups without any evidence of CMS . The mechanisms leading to progression from PMCV infection to CMS are currently unclear . CMS is present in Ireland. The first recorded outbreak of CMS occurred in 2012, associated with low-level mortalities over a period of 4–5 weeks followed by increased mortalities during bath treatment for sea lice . CMS is not a notifiable disease in Ireland, and there are no systematic records of its occurrence. Nonetheless, anecdotal information from field veterinarians and farmers suggest that CMS occurrence has steadily increased over the years. A retrospective study was recently conducted, using real-time RT-PCR with archived broodstock samples dating back to 2006, which suggests that PMCV may have been introduced into Ireland in two different waves, both from the southern part of the range for PMCV in Norway . PMCV was found to be largely homogenous in Irish samples, with limited genetic diversity. Further, the majority of PMCV strains had been sequenced from fish that were not exhibiting any clinical signs of CMS, which suggests possible changes in agent virulent and/or the development of immunity in Irish farmed Atlantic salmon . This paper describes the use of data-driven network modeling as a framework to evaluate the transmission of PMCV in the Irish farmed Atlantic salmon population and the impact of targeted intervention strategies. This approach can be used to inform control policies for PMCV in Ireland, as well as other infectious diseases in the future.The characteristics of the Irish Atlantic salmon have been described elsewhere , but briefly, in broodstock farms, eggs, and milt are obtained from sexually mature fish to produce fertilized eggs. In freshwater farms, fertilized eggs hatch, and fish are kept until smoltification, the stage where fish are ready to transition into the ocean . Some companies move the fish to net pens in freshwater lakes for the smoltification to occur there. After smoltification, fish are transported to seawater net pens, where they will grow until market size , with the possibility of being moved to other sea sites in between. Some of these fish are selected to become the broodstock for the next production cycle. These fish are transported from sea sites into freshwater broodstock facilities in late summer and early fall, to be stripped later in winter.These methods are based on those described by Widgren et al. . A stochastic within-farm model, linked to other farms through fish movements and local spread, was used to model the dynamics of PMCV infection in each farm. We developed a SIE compartment model with two disease states, susceptible and infected , grow benches and an environmental compartment, E. It was assumed that infected fish do not recover , and that susceptible fish can become infected through contact with PMCV present in the environment or via introduction of infected fish. To evaluate age-related differences in both the dynamics of PMCV infection within the host and in the likelihood of being moved , the two disease states, S and I, were further subdivided into three different age categories, indexed using j, including a. “egg-juveniles” from egg until 7 days prior to transfer to a marine farm, b. “smolt” from 7 days prior to the transfer to a marine farm to 180 days after the transfer, and c. “growthrepro” being more than 180 days in a marine farm. In those situations where this level of detail was not available for a particular fish group, all fish moving between freshwater farms were assumed to be egg-juveniles, and all moving between marine farms were assumed to be growth-repro. Therefore, Si,j , Ii,j , and Ei represents the six disease compartments and the environmental compartment within each farm i . A continuous-timediscrete-state Markov process with the Gillespie’s direct method, as implemented by Bauer et al. , was used to model the state transitions between the susceptible and infected compartments within each farm .
Model parameters were estimated from a previous study, which had been conducted in 2016 and 2017 to determine the prevalence of PMCV infection in Irish salmon farms by real-time RT-PCR . The sampling strategy was replicated to ascertain the status that could have been found if simulated farms had been sampled. In this study, sample collection was conducted on 22 farms from 30 May 2016 to 19 December 2017. A ranching farm is a freshwater broodstock farm that releases juvenile fish to the environment for conservation purposes. Some farms were sampled more than once over the course of the study, with the median samplings per farm in this group being 3.5 . A total of 1,201 fish were sampled during the study. Samples consisted of heart tissue across all fish age classes and ova. In this study, PMCV was detected at a low level in most sites, with only one clinical case of CMS occurring during the study period. We simulated sampling at each time point by randomly sampling fish within each farmand age category, as in the observed data set, from the number of susceptible and infected individuals at the time for the sampling point in the simulated farms. The aforementioned observational study also looked for PMCV in archived samples of Atlantic salmon broodstock from 2006 to 2016, seeking to determine whether the agent had been present in the country prior to the first case report in 2012 . For this, archived samples of broodstock Atlantic salmon were tested for each year from 2006 through 2016, using 60 archived pools per year. Samples are collected on an annual basis as part of the national disease surveillance programme, and consist of pooled organ homogenate supernatants which are then stored at −80◦C. In these samples, PMCV was first detected from broodstock fish on a marine site in July 2009, with infection in 100% of the pools . It was later detected in December of the same year in 100% of pools of broodstock fish at a second broodstock farm. These are subsequently referred to as the index cases, to be used as the starting point for simulation of the epidemic. The rationale behind setting these two farms as index cases is that they are the earliest detections , and do not seem to be epidemiologically connected. There were 9 parameters in the SIE compartment model . Following the approach used by Widgren et al. , and for model parsimony, the shed rate α was fixed at 1.0 per day, thereby defining the unit of the environmental infectious pressure variable ϕi. In the absence of more detailed data, it was assumed that the threshold for two seawater farms being connected by local spread was an euclidean distance of 10 km. It was further assumed that freshwater farms were not connected via local spread but only via fish movement.In the model, four event types were defined. “Enter” concerns hatchings and international imports. “Internal transfer” occurs on the day that individual fish change their age category, from egg-juvenile to smolt, or from smolt to growth-repro. “External transfer” occurs when fish move from one farm to another. “Exit” is linked with slaughter, euthanasia or international export, and from this point these fish are no longer included in the simulation. Each of the scheduled events was executed in the model once the simulation, in continuous time, reached the time for any of the events.