Utilizing spatial statistical models for implementing agronomic trials
Systems and methods for utilizing a spatial statistical model to maximize efficacy in performing trials on agronomic fields are disclosed herein. In an embodiment, a system receives first yield data for a first portion of an agronomic field having received a first treatment, and second yield data for a second portion of the agronomic field having received a second treatment different than the first treatment. The system uses a spatial statistical model and the first yield data to compute a yield value for the second portion of the agronomic field, where the yield value indicates an agronomic yield for the second portion of the agronomic field if the second portion of the agronomic field had received the first treatment instead of the second treatment. Based on the computed yield value and the second yield data, the system selects the second treatment and generates a prescription map including the second treatment.
1. A system comprising:
one or more processors;
a memory storing instructions which, when executed by the one or more processors, causes performing:
receiving first yield data for a first portion of an agronomic field, the first portion of the agronomic field subjected to a first treatment;
receiving second yield data for a second portion of the agronomic field, the second portion of the agronomic field subjected to a second treatment that is different than the first treatment;
using a spatial statistical model, computing a yield value, per location, for the second portion of the agronomic field based on a vector of covariate(s) and the first yield data, the vector of covariate(s) indicative of one or more measurable properties of the agronomic field per location, the yield value indicating an agronomic yield for the second portion of the agronomic field had the second portion of the agronomic field been subject to the first treatment;
determining, per location of the second portion of the agronomic field, a difference between the computed yield value and the second yield data;
selecting the second treatment based on the difference relative to a threshold; and
generating a prescription map based on the spatial statistical model, the prescription map including the second treatment applied to a portion of a field, which is larger than the second portion of the agronomic field.
2. The system of claim 1 , wherein the first treatment includes a first seeding population; and
wherein the second treatment includes a second seeding population different than the first seeding population.
3. The system of claim 1 , wherein the second treatment includes application of a pesticide or a nutrient; and
wherein the first treatment include no application or application of a different pesticide or nutrient.
4. The system of claim 1 , wherein the first treatment includes a first hybrid type; and
wherein the second treatment includes a second hybrid type different than the first hybrid type.
5. The system of claim 1 , wherein the spatial statistical model is configured to compute yield values as a function of a spatially correlated Gaussian process; and
wherein the spatial statistical model is further based on a mean yield for the agronomic field.
6. The system of claim 5 , wherein computing a yield value, per location, for the second portion of the agronomic field is based on:
y ( s i )=μ+ x i β+w ( s i )+ε i ,
where μ is the mean yield for the agronomic field, x i is the vector of covariate(s), β is a parameter vector, w(s i ) is a spatially correlated process, and ε i is an error process to be fitted based on variances in the agronomic field.
7. The system of claim 1 , wherein the one or more measurable properties include one or more of a percentage of organic matter, pH, cation-exchange capacity, elevation, soil type, and/or nutrient levels.
8. The system of claim 1 , wherein the one or more measurable properties include a vegetative index value for each of the locations.
9. The system of claim 1 , wherein selecting the second treatment comprises:
computing the threshold based on a standard deviation of the computed yield values; and
determining that a yield of the second yield data is greater than the computed yield value and, in response, selecting the second treatment.
10. The system of claim 9 , wherein each location is an equal size; and
wherein the vector of covariate(s) defines the one or more measurable properties for each of the locations.
11. A computer-implemented method comprising:
receiving first yield data for a first portion of an agronomic field, the first portion of the agronomic field subjected to a first treatment;
receiving second yield data for a second portion of the agronomic field, the second portion of the agronomic field subjected to a second treatment that is different than the first treatment;
using a spatial statistical model, computing a yield value, per location, for the second portion of the agronomic field based on a vector of covariate(s) and the first yield data, the vector of covariate(s) indicative of one or more measurable properties of the agronomic field per location, the yield value indicating an agronomic yield for the second portion of the agronomic field had the second portion of the agronomic field been subject to the first treatment;
determining, per location of the second portion of the agronomic field, a difference between the computed yield value and the second yield data;
selecting the second treatment based on the difference relative to a threshold; and
generating a prescription map based on the spatial statistical model, the prescription map including the second treatment applied to a portion of a field, which is larger than the second portion of the agronomic field.
12. The computer-implemented method of claim 11 , wherein the first treatment includes a first seeding population; and
wherein the second treatment includes a second seeding population different than the first seeding population.
13. The computer-implemented method of claim 11 , wherein the second treatment includes application of a pesticide or a nutrient; and
wherein the first treatment include no application or application of a different pesticide or nutrient.
14. The computer-implemented method of claim 11 , wherein the first treatment includes a first hybrid type; and
wherein the second treatment includes a second hybrid type different than the first hybrid type.
15. The computer-implemented method of claim 11 , wherein the spatial statistical model is configured to compute yield values as a function of a spatially correlated Gaussian process; and
wherein the statistical model is further based on a mean yield for the agronomic field.
16. The computer-implemented method of claim 15 , wherein computing a yield value, per location, for the second portion of the agronomic field is based on:
y ( s i )=μ+ x i β+w ( s i )+ε i ,
where μ is the mean yield for the agronomic field, x i is the vector of covariate(s), β is a parameter vector, w(s i ) is a spatially correlated process, and ε i is an error process to be fitted based on variances in the agronomic field.
17. The computer-implemented method of claim 11 , wherein the one or more measurable properties include one or more of a percentage of organic matter, pH, cation-exchange capacity, elevation, soil type, and/or nutrient levels.
18. The computer-implemented method of claim 11 , wherein the one or more measurable properties include a vegetative index value for each of the locations.
19. The computer-implemented method of claim 11 , wherein selecting the second treatment includes:
computing the threshold based on standard deviation of the computed yield values; and
determining that a yield of the second yield data is greater than the computed yield value and, in response, selecting the second treatment.
20. The computer-implemented method of claim 19 , wherein each location is an equal size; and
wherein the vector of covariate(s) defines the one or more measurable properties for each of the locations.