Brett Whelan

Future Directions of Precision Agriculture

Precision Agriculture is advancing but not as fast as predicted 5 years ago. The development of proper decision-support systems for implementing precision decisions remains a major stumbling block to adoption. Other critical research issues are discussed, namely, insufficient recognition of temporal variation, lack of whole-farm focus, crop quality assessment methods, product tracking and environmental auditing. A generic research programme for precision agriculture is presented. A typology of agriculture countries is introduced and the potential of each type for precision agriculture discussed.

The “Null Hypothesis” of Precision Agriculture Management

As precision agriculture strives to improve the management of agricultural industries, the importance of scientific validation must not be forgotten. Eventually, the improvement that is imparted by precision agriculture management must be considered in terms of profitability and environmental impact (both short and long term). As one form of precision agriculture, we consider site-specific crop management to be defined as: “Matching resource application and agronomic practices with soil and crop requirements as they vary in space and time within a field.” While the technological tools associated with precision agriculture may be most obvious, the fundamental concept will stand or fall on the basis of scientific experimentation and assessment. Crucial then to scientifically validating the concept of site-specific crop management is the proposal and testing of the null hypothesis of precision agriculture, i.e. “Given the large temporal variation evident in crop yield relative to the scale of a single field, then the optimal risk aversion strategy is uniform management.” The spatial and temporal variability of important crop and soil parameters is considered and their quantification for a crop field is shown to be important to subsequent experimentation and agronomic management. The philosophy of precision agriculture is explored and experimental designs for Precision agriculture are presented that can be employed in attempts to refute the proposed null hypothesis.

A preliminary approach to assessing the opportunity for site-specific crop management in a field, using yield monitor data

Abstract This paper proposes an Opportunity Index (Oi) for site-specific crop management (SSCM). In contrast to the traditional practice of uniform agronomic management, SSCM aims to match controllable inputs with spatially variable crop requirements. Farmers, however, are often left wondering how their yield maps can be used to justify a change to SSCM. The Oi is a single number that may be used in the process of this justification. The Oi is based on three components: (1) the magnitude of variation present in a yield map, relative to a certain threshold; (2) the average area within which yield is autocorrelated, relative to the minimum area within which variable-rate controllers (which physically implement SSCM) can reliably operate; and, (3) the economic and environmental benefit of SSCM relative to uniform management. The Oi was calculated for 20 Australian cropping fields and compared with b′ [Journal of Agricultural Sciences, Cambridge, 28 (1938)], which is an alternative method of quantifying the management opportunity from yield variation. A weak negative correlation was found to exist. Results suggest that a good opportunity for site-specific crop management exists when the Oi is greater than 20, although this is only a tentative recommendation.

Comparing temperature correction models for soil electrical conductivity measurement

There are various factors that affect soil electrical conductivity (EC) measurements, including soil texture, soil water content, cation exchange capacity (CEC) and others. Temperature is an important environmental variable, and different models can be used to correct for its effect on EC measurements and standardize the measurements to 25°C. It is relevant to analyze these models and to determine whether they are consistent with each other. Some models were wrongly cited. We found that the exponential model of Sheets and Hendrickx as corrected by Corwin and Lesch in 2005 performs the best. The ratio model also performs well between 3°C and 47°C.

Measuring the quality of digital soil maps using information criteria

One of the purposes of a soil map is to present information regarding the spatial variation of soil to the end user. In the past, statistical methods of uncertainty have been used to indicate map quality. In this paper, the information content of a map is proposed as a more suitable measure of map quality. The theory has its basis in information theory, more particularly Shannons information criterion. A modification of Shannons information criterion is presented. Other than being used as an indicator of map quality, the method is particularly useful as an aid during the map production process, for example, choosing block size or grid spacing. Four examples illustrate the concept and its potential use in soil science.

Sampling strategy is important for producing weed maps: a case study using kriging

Abstract Weed maps are typically produced from data sampled at discrete intervals on a regular grid. Errors are expected to occur as data are sampled at increasingly coarse scales. To demonstrate the potential effect of sampling strategy on the quality of weed maps, we analyzed a data set comprising the counts of capeweed in 225,000 quadrats completely covering a 0.9-ha area. The data were subsampled at different grid spacings, quadrat sizes, and starting points and were then used to produce maps by kriging. Spacings of 10 m were found to overestimate the geostatistical range by 100% and missed details apparently resulting from the spraying equipment. Some evidence was found supporting the rule of thumb that surveys should be conducted at a spacing of about half the scale of interest. Quadrat size had less effect than spacing on the map quality. At wider spacings the starting position of the sample grid had a considerable effect on the qualities of the maps but not on the estimated geostatistical range. Continued use of arbitrary survey designs is likely to miss the information of interest to biologists and may possibly produce maps inappropriate to spray application technology. Nomenclature: Capeweed, Arctotheca calendula (L.) Levyns.

Site-specific variation in wheat grain protein concentration and wheat grain yield measured on an Australian farm using harvester-mounted on-the-go sensors.

Accurately measuring and understanding the fine-scale relationship between wheat grain yield (GY) and the concomitant grain protein concentration (GPC) should provide valuable information to improve the management of nitrogen inputs. Here, GPC and GY were monitored on-harvester for three seasons across 27 paddocks on an Australian farming enterprise using two independent, on-the-go sensing systems. A Zeltex Accuharvest measured GPC (%) and a John Deere GreenStar system measured GY (t/ha). Local calibration in each season for Australian spring wheat significantly improved the prediction accuracy, precision, and bias of the Zeltex Accuharvest when compared with the initial factory calibration. Substantial variation in GPC and GY was recorded at the field scale, with the least variation recorded in both parameters in the wetter season. GY (CV = 38%) was twice as variable on average as GPC (CV = 19%) across the enterprise. At this enterprise scale, a negative correlation between GPC and GY was observed for a composite of the field data from all seasons (r = –0.48); however, at the within-field scale the relationship was shown to vary from positive (max. = +0.41) to negative (min. = –0.65). Spatial variation in GPC and GY at the within-field scale was described best in the majority of cases by an exponential semivariogram model. Within-field spatial variability in GPC is more strongly autocorrelated than GY but on average they share a similar autocorrelated range (a′ = ~190 m). This spatial variability in GPC and GY gave rise to local spatial variation in the correlation between GPC and GY, with 85% of the fields registering regions of significant negative correlations (P < 0.01) and significant positive correlations observed in 70% of fields. The spatial pattern in these regions of significantly different correlations is shown to display spatial coherence from which inferences regarding the relative availability of soil nitrogen and moisture are suggested. The results point to the suitability of these on-the-go sensors for use in more sophisticated agronomic and environmentally targeted nitrogen-use analysis.

A Parametric Transfer Function for Grain-Flow Within a Conventional Combine Harvester

Grain yield monitoring is an integral tool in the Precision Agriculture management system. When used in conjunction with a satellite-based navigation system, it provides spatial information on output variability, output response to managed inputs and is used to identify limiting resources in the crop production process. Accurately matching measured yield quantities with spatial units within a field is therefore important. At present, a simple linear time shift is employed by all commercial monitoring systems to account for the delay between GPS recorded positions and subsequent yield measurements. This study examines the internal process of grain transport to the sensor by monitoring the flow of strategically coloured grain. The flow is shown to be significantly influenced by mixing induced by threshing and auger transport processes. In contrast to the common assumption that grain moves as a spatially related cohort through to the sensor, the results suggest that a diffusion process is more realistic. A parametric model for the diffusion process is provided which suggests that from each individual yield measurement a maximum 20% of the mass could be assigned to a single spatial unit of the size that is typically allocated. The results imply that for further analyses, the inconclusive spatial origin and artificially smoothed quantities of instantaneous yield measurements should be considered.

Mapping almond orchard canopy volume, flowers, fruit and yield using lidar and vision sensors

Abstract This paper present a mobile terrestrial scanning system for almond orchards, that is able to efficiently map flower and fruit distributions and to estimate and predict yield for individual trees. A mobile robotic ground vehicle scans the orchard while logging data from on-board lidar and camera sensors. An automated software pipeline processes the data offline, to produce a 3D map of the orchard and to automatically detect each tree within that map, including correct associations for the same trees seen on prior occasions. Colour images are also associated to each tree, leading to a database of images and canopy models, at different times throughout the season and spanning multiple years. A canopy volume measure is derived from the 3D models, and classification is performed on the images to estimate flower and fruit density. These measures were compared to individual tree harvest weights to assess the relationship to yield. A block of approximately 580 trees was scanned at peak bloom, fruit-set and just before harvest for two subsequent years, with up to 50 trees individually harvested for comparison. Lidar canopy volume had the strongest linear relationship to yield with R 2 = 0.77 for 39 tree samples spanning two years. An additional experiment was performed using hand-held photography and image processing to measure fruit density, which exhibited similar performance ( R 2 = 0.71 ). Flower density measurements were not strongly related to yield, however, the maps show clear differentiation between almond varieties and may be useful for other studies.

The Analysis of Spatial Experiments

Anyone with an interest in precision agriculture has already formed a hypothesis that the field is a sub-optimum management unit for cropping. The role of experimentation is to test this hypothesis. Geostatistics can play an important role in analysing experiments for site-specific crop management: put simply, spatial autocorrelation must be accounted for if one is to draw valid inferences. We provide here some background to the basic concepts of agronomic experimentation. We then consider two broad classes of experimental design for precision agriculture (management-class experiments and local-response experiments), and show, with the aid of case studies, how each may be analysed geostatistically. Ultimately though, if farmers are compelled to use relatively simple designs and less formal analyses, then researchers must follow and adapt their geostatistical analyses accordingly.