Better Fertilizer Decisions for Crops: Developments of a National Web-Based Database to Support Soil Test Calibration.

Australia, like many countries, has carried out numerous fertilizer response experiments with soil tests but until recently the data were fragmented and inaccessible to crop nutrition advisers and consultants. The Better Fertilizer Decisions for Crops (BFDC) project developed and now curates a national database of 5662 trial treatment series undertaken at 2724 sites. These span five decades of 1780 N, 2380 P, 365 K and 286 S experiments with soil test values and crop responses on cereal, pulse and oilseed crops. Minimum data requirements from past, present and future experiments for inclusion are soil type, a recognized soil test and that an estimate of crop yield with nil fertilizer and the maximum yield could be obtained.

This database is used on-line by >350 trained and authorized users to obtain soil test-crop response relationships for the interpretation of soil test results using an agreed and statistically valid method. Users can filter the data by attributes that include crop type, soil type, soil test, grain yield and growing season rainfall. Nutrient response criteria are obtained by fitting response curves from which critical soil test values and confidence limits are determined for 80%, 90% and 95% of relative yield. These results are presented with the critical values and ranges for each of these relative yields.

The second phase of BFDC has three enhancements. An on-line registration and training module has been developed to enable users, including university students, to best use the database interrogation tool. Registered users are now able to enter experimental data and once checked these are added to the general data pool. Finally, a protocol has been developed to enable data from long-term fertilizer experiments to be included and analysed.

This project has enabled data to be preserved, accessed, interpreted and added to, so supporting the evidence based approach to soil test interpretation. The authors commend this approach to other precincts where current and legacy data awaits capture and systematic interrogation, and to identify knowledge gaps to guide future research.