The DGT technology was tested across all of the major South Australian potato growing regions by establishing a total of 15 replicated trials and 9 demonstrations (Table 1), encompassing a wide range of soil and potato types, including fresh, processing, crisping and seed varieties.

Table 1. Location and timing of replicated and demonstration trials across South Australia.
Location Production type Replicated trial Timing Demonstration trial Timing
Kangaroo Island Seed 3 1 x summer 2014
2 x summer 2015
3 1 x summer 2014
1 x summer 2015
1 x winter 2015
Riverland Fresh 2 1 x summer 2015
1 x autumn 2015
Mallee Fresh and seed 4 1 x summer 2015
3 x autumn 2015
Adelaide Plains Fresh  – 2 Autumn 2015
Lower Lakes Crisping 1 1 x winter 2015
South East Processing 5 5 x spring 2015 4 3 x spring 2014
1 x summer 2015
15 TOTAL 9


The yield response trials involved applying five rates of P fertiliser (including 0 kg P/ha) in a randomised replicated trial design to assess the effect on potato yield. Soil P status was assessed upon trial commencement using the common Colwell, Olsen and Bray 2 methods and Phosphorus Buffering Index and the DGT P test were also used to quantify the accuracy of the DGT test in comparison to the historically used soil P tests.

Field experiments

All field experiments were conducted on privately operated potato growing farms, embedded within commercial crops grown under centre pivot or solid set irrigation for seed, fresh, processing or crisping purposes. For the replicated trials, fifteen randomised complete block experiments were conducted with five P treatment rates applied, replicated thrice. Individual plot size was commonly 6 mounds wide by 8 m long (approximately 41 m2 per plot). For demonstration sites, three plots were pegged within a production area, with plots 1 and 3 receiving the P application in use by the farmer and plot 2 receiving no P fertiliser. Soil sampling was conducted by collecting a minimum of 30 samples from each treatment plot to a depth of 15 cm.

Total fertiliser requirements for each plot were calculated to match the growers’ application practices (including nitrogen, potassium, sulphur and trace elements), with only the P content being manipulated to match the treatment rates required. P fertiliser in the form of triple super phosphate was used on Kangaroo Island and MAP was used for all other trial sites. All fertiliser was broadcast uniformly by hand on the surface of the plots prior to planting; in some cases a single pass shallow cultivation then followed. The crop was planted with commercial seeding equipment in use by the individual farmer.

Harvest was conducted following plant senescence by digging 3 x 2 m of mounds (by hand) at random within each plot (ensuring buffer zones were maintained); a total of 6 m of mound were harvested per plot. All tubers present were collected, graded into the size categories in use by the grower and the total number and weight of tubers was recorded for each category.

Analytical methods

Soil samples were sent to APAL Agricultural Laboratory where they were dried to 40 degrees Celsius and ground to <2 mm. Samples were analysed for pH (water and CaCl2), EC1:5, OC (Walkley Black), phosphorous concentration (Colwell, Olsen, Bray 2P), phosphorous buffering index (PBI), potassium (Colwell), ammonium N, nitrate N, extractable sulphur (KCl), exchangeable cations (1 M ammonium acetate) and DTPA trace elements following the methods of Rayment and Lyons (2011). Diffuse gradients in thin films (DGT) phosphorous concentration was also determined following the method of Mason et al 2010.

Statistical method

Analysis of variance (ANOVA) was conducted on the total potato tuber yields x P treatment rate for each of the trial sites. Plot position was used conventionally as a covariate to adjust for within-site fertility trends; where there was no trend apparent, the conventional randomized block ANOVA was used. Yield data were plotted and quadratic trend lines fitted to determine maximum yield (Ymax) and yield at no applied P (Y0P). Yield penalty to no applied P fertiliser was expressed in relation to: (Y0P/Ymax) x 100 and is referred to herein as Relative Yield %. RY was subsequently correlated with the soil test data for each P measurement method and Mitscherlich regression models fitted to calculate the critical P value to obtain 90% RY.

Project Partners

Potatoes South Australia PIRSA RSSA National Landcare Programme The University of Adelaide premiumgreen-logo