Summary

Agricultural Production Systems sIMulator (APSIM) modelling identified the optimal flowering period (OFP) at Palmer in the Murray Plains region based on climatic data from 1973-2023.

Wheat OFP is between the 25th August and 8th September. The optimal sowing date range for current elite wheat varieties varied from 2nd April to 19th May.

Barley OFP is between the 2nd and 26th August. Very quick spring barley varieties have optimal sowing range of 12th to 23rd April in this region; significantly earlier than the same wheat phenology.

Conditions under which drone technology could most accurately estimate plant establishment and biomass were also demonstrated.

Background

Understanding optimal sowing dates and flowering periods for wheat and barley varieties regionally helps growers balance the risk of frost, heat and drought stress to maximise grain yield. The optimal flowering periods developed in SAGIT project (SUA1021) required further validation with another growing season of field trial data in the Murray Plains. The potential of early sowing slower maturing wheat and barley varieties in this region also required further investigation.

Drone technology has the potential to be used to automate laborious plant counts for cereals. However, further research is needed to develop and validate a method to increase accuracy of counts. Drones also have the potential to gather non-destructive estimates of plant biomass using different sensors, validated using plants with different plant growth habits.

Research Aims

The core objectives of the project were to:

• Refine sowing times and the OFP for wheat and barley.
• Examine the potential for early sowing of slow maturing wheat and barley.
• Identify growth traits to improve drone biomass and growth modelling in wheat and barley using drone imaging.
• Develop drone imaging method to accurately count plant establishment.

In The Field

APSIM modelling was undertaken to determine the OFP and optimal sowing date for wheat and barley on the Murray Plains and validated with a time of sowing by variety field trial.

Drone imagery was collected over the field trials to develop a thresholding method for estimating plant establishment. Varieties with different architecture and phenology were examined to fine tune biomass estimates using drones fitted with Red, Green and Blue (RGB) or multispectral sensors.

Results

The 2023 growing season was drier and warmer than the two seasons in the previous SAGIT investment. These conditions meant that flowering dates, particularly from April germination, were much earlier permitting improved validation of the start of the OFP.

APSIM modelling indicated the OFP for wheat at Palmer is between 25th August and 8th September. The 2023 field trial helped validate this, with maximum grain yield coming with a flowering date of the 2nd September. The range in optimal sowing dates for current elite wheat varieties varied from 2nd April to 19th May. For early sowing, the mid to slow spring wheats (Rockstar and Denison) had more stable and higher grain yields across growing seasons than the winter wheat, but new quick winter variety Mowhawk showed promise.

Optimal sowing dates were modelled for different spring wheat phenologies:

• Quick (Vixen, Anvil CL Plus): 6th May-19th May
• Quick-mid (Ballista, Calibre, Scepter): 30th April – 13th May
• Mid-slow (Rockstar, Sherrif CL Plus): 24th April – 11th May
• Slow (Denison, Valiant CL Plus): 2nd April^* – 29th April

^*This date was based on extrapolation so should be interpreted with caution.

APSIM modelling indicated the OFP for barley at Palmer is between the 2nd and 26th August. Current quick spring barley varieties have an optimal sowing date ranging from 12th April to 23rd April in the Murray Plains region; significantly earlier than the same wheat phenology indicating that higher grain yields are achieved with earlier flowering.

A mean accuracy of 88 per cent was achieved in the first year of plant counts extracted from drone imagery with individual seedlings identified using still images across each plot at the 1-2 leaf stage. A thresholding method was used to determine individual plants from green pixels and permitted direct plant counting. Lower sowing densities reduced count accuracy. Reducing drone altitude from 10 metres to 5 metres increased accuracy without increasing blur from wind turbulence.

Drone imagery was able to detect plant height differences directly related to growth stage at any imaging time point, correlating well with manual biomass cuts. The two model barley varieties Combat and Maximus CL had very strong correlations between plant height and manual dry matter in 2023. The differences in barley architecture (prostrate vs erect plant types) only explained a smaller component of the variation, with other potential drone indices tested not improving accuracy compared to manual calibration biomass cuts. The plant height correlation was consistently strong across the whole growing season, especially compared to NDVI.