Short-term associations between ambient ammonia concentrations and growing-finishing pig performance and health.

Using sensors, the health and welfare of growing-finishing pigs can be continuously monitored by detecting deviations from pigs' normal behaviour, but the validity of such algorithms requires improvement. As changes in the environment influence pig behaviour, monitoring temporal changes in environmental factors may help identify periods with a higher risk of welfare issues. The real-time relationships between pig welfare and many environmental factors are, however, not well-understood. This study examined the short-term associations of ambient ammonia with indicators of pig production and health. Ambient ammonia concentrations were monitored with sensors during the growing-finishing period of three rounds at a German (farm G, n = 110 pigs/round) and one round at a Danish farm (farm D, n = 144 pigs). Body weight was estimated daily using 3D cameras (both farms), feed intake was recorded using electronic feeding stations (only farm G), and health indicators were recorded during twice- (farm G) or thrice-weekly (farm D) farm visits. Using splines (generalised additive models), ammonia concentrations were linked to indicators of pig production and health in real time and, for body weight, at a lag of 1, 2, 3 and 7 d. We found a range of relationships between ambient ammonia (5 - 50 ppm) and production or health indicators (i.e. body weight (real-time and lagged), feed intake, coughing, sneezing, pen fouling, diarrhoea, and tear staining), but they were highly inconsistent between farms and pig rounds. Part of this inconsistency may be due to differences in manure management and sensor locations, or could be explained by age or seasonal effects (e.g. heat stress). More robust relationships were identified for clinical measures related to pig behaviour, where tail damage and skin lesions linearly increased with ammonia from low concentrations (5 - 10 ppm) onwards, hence suggesting more tail biting and aggression at higher ammonia concentrations. In conclusion, ambient ammonia did not clearly associate with pig performance and health in the short term, while higher ammonia concentrations were related to higher occurrences of clinical signs reflecting undesirable behaviours. Therefore, daily ammonia measurements using sensors may be of limited value in identifying health issues in pigs, but they may aid in detecting periods with high risk of aggressive or tail biting behaviours that require interventions. As ambient ammonia was confounded with other environmental measurements, such as ambient temperature or carbon dioxide concentrations, identified associations should be interpreted cautiously or with ammonia as general indicator of air quality.