Author
 |
Acock, Basil |
|
PACHEPSKY, YAKOV - DUKE UNIVERSITY |
|
Submitted to: European Conference on Precision Agriculture Proceedings
Publication Type: Proceedings Publication Acceptance Date: 4/11/1997 Publication Date: N/A Citation: N/A Interpretive Summary: Using recent advances in global position sensors, farmers can now determine exactly where they are in a field. With this knowledge, and monitors that measure yield continuously while the crop is being harvested, they can prepare yield maps for their fields. Other advances in engineering enable them to record certain soil and plant properties on the move, and to apply spatially-variable amounts of various agricultural inputs. The use of these technologies is called precision farming, and its purpose is to manage land units smaller than a field to maximize profitability and minimize environmental pollution. The major problem with precision farming is that we do not yet understand the reasons for yield variability or know how to "prescribe" different management treatments for various parts of the field. Most researchers in this area are measuring soil nutrients in a grid pattern over the field, assuming that nutrient imbalance is responsible for the yield variation. They are failing to find any consistent correlation between nutrients and yield. We will be able to understand yield variability only when we consider all the major factors limiting plant growth. The only practical way of considering all relevant factors day by day through the season is to use mechanistic crop models: models that attempt to simulate the principal mechanisms in the soil/plant/atmosphere system occupied by the crop. This paper presents the rationale and techniques for using mechanistic crop models in precision farming. Technical Abstract: Precision farming employs many advances in engineering to determine location within a field, record soil and plant properties including yield, and to apply spatially-variable amounts of various agricultural inputs. Most researchers in this area are measuring soil nutrients in a grid pattern over the field, sometimes overlaying nutrient treatments, assuming that nutrient imbalance is responsible for the yield variation. They are failing to find any consistent correlation between nutrients and yield. This is the principal hole in Precision Farming - a lack of understanding of the reasons for yield variation. We will be able to prescribe treatments for yield variability only when we understand the reasons for that variability, and we will understand only when we consider all the major factors limiting plant growth. The only practical way of considering all these factors is to use mechanistic crop models: models that attempt to simulate the principal mechanisms in the soil/plant/atmosphere system occupied by the crop. However, because of limitations in our knowledge of plant behavior, mechanistic crop models are still imperfect tools. It will be necessary to validate and possibly calibrate each crop model for various points in the field before using it in an iterative procedure to simulate and understand the pattern of yield variability observed. Even when we do understand all the reasons for yield variability, any prescription will depend on weather pattern. Precision farming will only work when we can prescribe management action during the course of the growing season. For these reasons, mechanistic crop models are essential to precision agriculture. Rather than discarding them because of their imperfections, we should continue to improve them. |
