Using weather data to predict the presence of Lucilia spp. on sheep farms in New Zealand

Flystrike remains an important animal health issue on New Zealand sheep farms. To date no useful predictive tool to assist farmers to develop control options has been available. The aim of this study was to use National Institute of Water and Atmospheric Research (NIWA) virtual climate station data in New Zealand to develop a weather-based model to accurately predict the presence of Lucilia spp. on sheep farms throughout New Zealand. Three LuciTrap® baited fly traps were positioned on each of eight sheep farms throughout New Zealand (5 in the North Island and 3 in the South Island). The traps were put out for both the 2018-2019 and 2019-2020 seasons. They were emptied each week and the flies morphologically identified; with the counts of Lucilia cuprina and L. sericata combined as Lucilia spp. The count data for Lucilia spp. for each week of trapping was transformed into a binary outcome and a generalised linear mixed effects models fitted to the data, with farm as a random effect. The dependent variable was Lucilia spp. flies caught, yes or no, and the independent variables were mean weekly climate variables from the nearest NIWA virtual climate station to that farm. The model was trained on the 2018-2019 catch data and tested on the 2019-2020 catch data. A cut point was identified which maximised the model's ability to correctly predict whether Lucilia spp. were present or not for the 2019-2020 catch data, and the sensitivity, specificity, accuracy, and area under the curve (AUC) of the model calculated. The final model included just 3 significant variables, mean weekly 10 cm soil temperature, mean weekly soil moisture index, and mean weekly wind speed at 10 m. Mean weekly 10 cm soil temperature accounted for 64.7% of the variance explained by the model, mean weekly soil moisture index 34.7% and mean weekly wind speed at 10 m only 0.6%. The results showed that the predictive model had a sensitivity of 0.93 (95% CI = 0.80-0.98) and a specificity of 0.75 (95% CI = 0.62-0.85), using a cut point for the probability of Lucilia spp. being present on farm = 0.383. This model provides New Zealand farmers with a tool which will allow them to know when Lucilia spp. flies will likely be present and thus more accurately plan their interventions to prevent flystrike.

Using a rule-based envelope model to predict the expansion of habitat suitability within New Zealand for the tick Haemaphysalis longicornis, with future projections based on two climate change scenarios

Haemaphysalis longicornis is the only species of tick present in New Zealand which infests livestock and is also the only competent vector for Theileria orientalis. Since 2012, New Zealand has suffered from an epidemic of infectious bovine anaemia associated with T. orientalis, an obligate intracellular protozoan parasite of cattle and buffaloes. The aim of this study was to predict the spatial distribution of habitat suitability of New Zealand for the tick H. longicornis using a simple rule-based climate envelope model, to validate the model against published data and use the validated model to project an expansion in habitat suitability for H. longicornis under two alternative climate change scenarios for the periods 2046-2065 and 2081-2100, relative to the climate of 1981-2010. A rule-based climate envelope model was developed based on the environmental requirements for off-host tick survival. The resulting model was validated against a maximum entropy environmental niche model of environmental suitability for T. orientalis transmission and against a H. longicornis occurrence map. Validation was completed using the I-similarity statistic and by linear regression. The H. longicornis climate envelope model predicted that 75% of cattle farms in the North Island, 3% of cattle farms in the South Island and 54% of cattle farms in New Zealand overall have habitats potentially suitable for the establishment of H. longicornis. The validation methods showed an acceptable level of agreement between the envelope model and published data. Both of the climate change scenarios, for each of the time periods, projected only slight to moderate increases in the average farm habitat suitability scores for all the South Island regions. However, only for the West Coast, Marlborough, Tasman, and Nelson regions did these increases in environmental suitability translate into an increased proportion of cattle farms with low or high H. longicornis habitat suitability. These results will have important implications for the geographical progression of Theileria-associated bovine anaemia (TABA) in New Zealand and will also be of interest to Haemaphysalis longicornis researchers in Australia, Japan, Korea and New Zealand.