Abstract/Summary

The purpose of the present study was, firstly, to examine current practice for the agronomy of grass-clover mixed swards used for silage-making in the UK, and secondly, to develop and validate a Near Infra-Red Reflectance Spectroscopy (NIRS) equation capable of predicting clover concentration (CC) in undried and unmilled grass-clover silage samples. A calibration set of 94 grass-clover (white, trifolium repens, and red, trifolium pratense) mixture silage samples were sourced from UK farms and an accompanying questionnaire was used to obtain information on the sward agronomy used to produce each sample. Questionnaire data highlighted that (i) reducing the use of fertiliser inputs (ii) increasing uptake of new varieties, and (iii) increasing the farmer’s ability to measure botanical composition as potential strategies for improving the utilisation of clover in grass swards. Botanical composition was measured by hand separation for each sample and a new NIRS equation was created and assessed using blind validation with an independent set of 30 grass-clover samples. The relative standard error of cross validation (SECV, as a percentage of the measured mean) of the optimised equation produced was 36.8%, and, in an independent validation test, the ratio of standard error of prediction to the standard deviation of the reference data set (RPD) was 1.56. The equation could be improved by increasing accuracy at high CCs but showed promise as a simple tool to assist growers in sward management decisions.