Modelling of strategic grass harvest management

Abstract

Grass harvest plays a crucial role in milk production. Farmers face the problem of timing the harvest with respect to quality (crude protein content) and quantity (dry matter yield). Literature suggests that harvesting more frequently and thereby keeping the grass short (light harvesting) will result in a higher crude protein content than for less frequent harvesting (heavy harvesting), but will also result in lower yield. However, this trade-off between quality and quantity is based on experiments with typically extremely light or heavy strategies, which are not realistic. We performed a model study to investigate the effect of harvesting strategy on quality and quantity under more realistic harvesting strategies. We used a potential crop growth model adapted to grass. A parameter calibration was carried out with yield and weather data of 2010 and 2011 from one specific field at Knowledge Transfer Centre De Marke (Netherlands), resulting in a correlation coefficient of R=0.879. Validation was performed using 13 years (2002-2014) of weather and yield data on the site (R=0.776). With this model, the yields of five different harvest scenarios were predicted, based on 22 years of weather data at the site. In contrast to literature sources, the model predicted, on average, around 10% more annual yield under the light harvest strategies. The difference may be explained by a large regrowth delay after a heavy harvest, caused by a heavy loss of leaf area. Consequently, potential growth days are utilized more efficiently under light harvesting. These predictions could mark the start of the development of new harvest strategies, based on a model-experiment cycle.