20 IREC Farmers' Newsletter No. 199 — Autumn 2018 Location district and yields There is significant diversity of season and soils within the southern growing region. The contributed data included fields from south of Jerilderie to as far west as Balranald and north of Griffith. Farms were grouped into their geographic districts or locations. The location grouping had the most significant effect on yield accounting for 23.6% of yield variation within the data (Table 1). The factors behind regional differences are numerous, and include climatic variation, predominant soil type, irrigation layout and design, farm size, cotton management experience and available machinery, to name a few. There were significant interactions between location and other yield drivers and these interactions reflected how districts vary in their farming practice, experience, farm infrastructure and design. As cotton is a tropical plant it could be assumed that higher yields would be associated with warmer districts, however in the 2016–17 season this did not hold absolutely true as many other factors had an influence with variability between districts. Sowing date The average date of sowing was much later in the 2016–17 season compared with previous crop years. This was a result of a very wet, cool winter that delayed picking of the previous crop, which in turn delayed some post-harvest field operations and didn’t allow time for adequate ground preparation for the subsequent crop. Some fields were able to be planted in late September to early October but these were few. The yield relationship to sowing date is very significant (Figure 2). Based on data for the 2016–17 season, and in spite of a delayed start to the season, the optimum sowing time was the second week of October. The predicted optimum sowing date for the season was 10–17 October. A 3% yield decline was predicted if sowing occured in the week either side of the optimum. The predicted decline for sowing in the second week either side of the optimum was 9% and in the third week, 18%. A lot of other factors affected yield in the 2016–17 season, however it is interesting to note that the predicted optimum time of sowing by CSIRO, using the OZCOTT model, was the week around 13 October (Australian Cotton Production Manual 2017, p 87). The interaction of sowing date with location was significant but complicated and no clear trends stand out for each location. More analysis over a number of years is required before strong conclusions for each location can be made. Field history and rotation For another angle of the analysis, fields were grouped according to previous crop history. If a field was cropped in the previous summer (e.g. with corn, cotton or soybean) it was classified as ‘back to back’ (B/B) in the analysis. If a field was fallow over the summer regardless of the previous winter crop then it was considered ‘fallow’. All summer crops were grouped as it was assumed that ground preparation may have been compromised with the wet winter compared with a dry bed preparation in the summer. Figure 2. Average cotton yield, as a result of sowing date, in the southern cotton growing region, 2016–17 Yield (bales/ha) Sowing date R2 = 0.41 Location in the southern growing region will have a significant influence on yield, reflecting a range of differences between districts.