Virtual Group - Real time estimatation of biomass

Summary

Accurately assessing feed-on-offer is a key component for the effective allocation of feed to different classes of stock.  Producers currently assess feed-on-offer visually, based on an estimate of average height and sward density.  Without clear reference criteria this becomes a subjective exercise and introduces errors into feed budgeting.

This project formed part of MLA's Producer Research Site program of the southern Feedbase Investment Plan. In particular, it supported the MLA-funded project B.GSM.0010 – Real-time pasture biomass estimation.

The potential for normalised differential vegetation indices (NDVI) to measure feed-on-offer was assessed at six high rainfall Victorian sites between 2014 and 2016.  Up to three assessments were made each year at each site.  At each harvest the NDVI of between 12 and 15 quadrats was measured using an active optical sensor (GreenSeeker®) and correlated against the herbage yield for the quadrat. 

In the second year, additional measurements were made with a CropCircle® sensor to examine the potential for different wavelengths to improve the correlation with feed-on-offer. The data obtained was analysed at UNE as a component of the B.GSM.001 project

Cooperating producers visually assessed feed-on-offer at the harvests, which was then compared to actual yields.

The correlations between NDVI and feed-on-offer varied between sites and seasonally and there was no one consistent calibration applicable to all situations.  Correlations improved when the height of the sward was included as a variable.  However, in the majority of harvests, height alone was superior to NDVI and in just over half of the situations, height was comparable to, or better than, the relationship fdp.between NDVI and height.
Different producers had different abilities to accurately assess feed-on-offer.  Some had a high level of consistency in their assessments, although they were often consistently above or below the actual figures.  Feed-on-offer was more likely to be overestimated at low feed availabilities and overestimated at high availabilities, particularly if senescent material was present.

In the absence of robust calibrations for NVDI at the trial sites and the reasonable correlations between feed-on-offer and sward height, greater emphasis should be placed on encouraging the use of rising plate meters for assessing feed-on-offer.

This project received additional collaborative support from Agriculture Victoria, allowing the group to increase the scope of the research undertaken.