Spatial and Temporal Dynamics of Alfalfa Production in Relation to Precision Agriculture and Integrated Crop Management Systems

Principle investigators:

Richard Leep, Roger Brook, Darryl Warncke, James DeYoung, Dennis Pennington, Peter Jeranyama, Tim Dietz, Bernie Knezek, Jim Bronson, Jerry Lindquist, and Bill Steenwyck

Abstract:

This research involves the study of spatial and temporal dynamics of alfalfa production and whether these dynamics can be managed through precision agriculture and integrated crop management systems.  Alfalfa is an important forage crop to Michigan dairy and livestock producers as it represents approximately 25% of the daily ration of dairy herds.  Soil fertility and insect management as well as variation in forage quality are factors, which can affect profitability of alfalfa producers.  When soil potassium levels are too high, alfalfa can take up too much potassium and when this alfalfa is fed to close-up cows in there is increased milk fever.  In addition, low soil pH often results in either poor stands or decreased stand life of alfalfa.  The first year of this project was dedicated to baseline data collection and evaluation.  Six fields were mapped for elevation, soil type information and partitioned them into smaller units for the study.  Three of the fields were already established while the other three fields were established in 2000.  Soil samples were analyzed on a one-acre grid. Forage quality was determined from hand clipped samples from the one-acre grid.  Yield monitors were installed and calibrated on hay mowers.  The yields from the monitors were compared with hand-harvested plots taken from both ¼ and one- acre grids.  In addition, alfalfa yield was calculated using normalized difference vegetation index (NDVI) from near infrared aerial photographs taken prior to the second and third harvest at the KBS alfalfa fields.   Potato leafhopper and alfalfa weevil populations were also sampled on a one-acre grid throughout the growing season to determine their spatial patterns within alfalfa fields. Results given below represent only a portion of the data analyzed.  Smaller replicated experiments have been established in each production field, which compares site-specific fertility management with whole field management based upon site specific or whole field soil test values.  The results below represent only a partial glimpse of the entire research data collected thus far.  However, the selected data shows significant findings from the research effort.

Objectives:

1)      Establishing the variability of soil properties (particularly pH and available K) that exist in typical Michigan alfalfa fields. 

2)      Determining what relationships exist in commercial alfalfa fields between alfalfa quality (yield and forage quality) and soil variables (as measured above).

3)     Determining the distribution of potato leafhopper populations and their effect on forage quality.

4)      Determine the most reliable and accurate method of estimating yield variation within alfalfa fields including yield monitors, near infra- red aerial photography analysis, and remote sensing from satellite image.

5)      Testing methods for reducing the variability in yield, quality and persistence in fields of alfalfa using precision agriculture sampling and treatment strategies. 

Results:

Potato leafhopper populations were mobile and asymmetrically distributed (Figure 1 and 2).  The high leafhopper population shifted from the SW corner of the field to the NE corner of the field in a 10-day period prior to the third harvest at KBS dairy farm.  The dynamics of population mobility and distribution were found in all six of the fields under study.  These patterns suggest a need to evaluate new control strategies for potato leafhopper, which will be further explored in the 2001-growing season.

Figure 1. Potato leafhopper distribution  at KBS on 7/27/00	Figure 2. Potato leafhopper distribution  at KBS on 8/7/00

Alfalfa weevil populations were also asymmetrically distributed the same field at the KBS dairy farm (Figure 3).  Alfalfa yield was negatively correlated (r = –0.35) with alfalfa weevil population in the first cutting at the KBS dairy farm (Figure 4).  Significantly higher alfalfa weevil populations were found in areas of the field at KBS, which were low in soil K at first cutting, thus, a negative correlation (r = -57) with soil K (Figure 5). 

Figure 3. Distribution of Alfalfa weevil at KBS on	Figure 4. First cutting yield of alfalfa at KBS on 5/23/00.

5/15/00

Total yield of alfalfa at KBS were significantly correlated with available soil potassium as well (r= 0.57).  Soil potassium levels and accumulated yield of alfalfa at KBS is given in Figures 5 and 6.  In addition, forage quality differences were significant within fields (data not shown).  These differences may be due to soil type and insect distribution.  Data is being further analyzed to determine the correlations between quality and other factors

Figure 5. Available soil K at the KBS dairy alfalfa field.	Figure 6.  Total dry matter yield in tons/acre for three cuttings at the KBS dairy alfalfa field in 2000.

Comparisons of yield using a yield monitor which was installed and calibrated in hay mowers with hand-harvested plots taken from grids and calculated yields using normalized difference vegetation index (NDVI) from near infra red aerial photographs taken prior to the second and third harvest at the KBS alfalfa fields were done in 2000.   The results of these methods are given in Figures 7, 8,9 and 10.  Both the yield monitor and NDVI show high and low yielding strips running North and South on the East side of the field.  The East side of the field grew conservation strips of corn and alfalfa prior to plowing and establishment of the field with alfalfa in 1998.  The low yielding strips represent areas where alfalfa was grown prior to the new seeding.        

Figure 7. Yield of third cutting using a yield monitor attached to mower-conditioner. (Yellow=low, green=high)	Figure 8. Yield of third cutting using (NDVI) from NIR aerial photograph. (Red=low, green=high)

Figure 9. Yield of third cutting using hand-harvested samples from one-acre grid in field.	Figure 10. Yield of third cutting using hand-harvested samples from 1/4 acre grid in field.

This would indicate a probable autotoxic effect of alfalfa being planted after alfalfa in the low yielding strips compared to higher yielding strips planted following corn.  It appears that even a ¼ acre grid hand sampling did not accurately represent the yield variation within the field as well as either the aerial near infra-red photograph or the yield monitor. 

First cutting yield of alfalfa was negatively correlated with weed growth (r = -0.74) in the first cutting at the Ionia location.  The weeds consisted mainly of annual broadleaf as to a lesser extent perennial grasses including quackgrass (Agropyron repens L.) and orchardgrass (Dactylis glomerata L).  The maps of weeds and first cutting yield from hand-harvested samples from the Ionia location are given in Figures 11 and 12.  Since this field was already established prior to our studies, we cannot determine what factors may have caused this spatial distribution of weeds in the field.  Studies initiated on three newly established fields in 2000 will be studied over the next several years to determine factors affecting alfalfa persistence and weed invasions into fields.

Figure 11. Map of weed infestation in first cutting at the Ionia dairy farm alfalfa field. (red=high, green=low)	Figure 12. Yield of first cutting at the Ionia dairy farm alfalfa field. (red=low, green=high)

The methods used and results of first year small replicated plots comparing site specific versus whole field management of boron based upon boron soil variation are given below.

Methods

Small plots were established within established alfalfa fields, which were two years old.  Plot size was 10 X 50 feet.  The fields were located in Kalamazoo County at the Kellogg Biological Station dairy farm (KBS1), in Ionia County (Ionia 1) and in Osceola County (Osceola 1).  Treatments consisted of boron application based upon site-specific soil test results for a plot compared to boron application based upon the mean soil test of the field.  Boron fertilizer was broadcast applied just after growth initiated in the spring of 2000.  Forage yield was taken by harvesting a 2 square foot area within the plots.  Forage quality including crude protein, acid detergent fiber and neutral detergent fiber, was determined using NIRS.

Results

At KBS1 site-specific management of boron was associated with a higher yield response compared with whole field management. Also without boron application in both site specific and whole field management did not result in similar yields to those of boron application (Table 1). All cuttings but the third cutting did not respond to boron management. In this case site-specific management at KBS 1 was associated with a higher yield compared with whole field management in the third cutting (Table 2).

At Osceola 1, there were no significant differences between whole field and site-specific management (Table 3). However, site-specific management in either without boron or with boron treatments had significantly higher yields than whole field management (Table 3). Total seasonal yields were higher within site-specific management of boron; however, individual cuttings were not different from each other (Table 3).

At Ionia site-specific management of boron produced significantly higher yields compared with whole field management in the first cutting (Table 4). However, in the second cutting, whole field management had significantly higher yields compared with site-specific management. Total season yields were not different among management and boron treatments (Table 4). In the second cutting, whole field management was associated with higher neutral detergent fiber (NDF) concentrations (Table 5). Also, in the second cutting within whole field management boron application was associated with a higher NDF concentration compared with without boron treatment (Table 5). In both site specific and whole field management, without boron treatment was associated with a higher crude protein (CP) concentration (Table 5). 

Table 1. Yield response to management and Boron applications at KBS1 in small plots.

Table 2. The effect of management and cut on alfalfa yield at KBS1 in small plots.

Table 3. Yield  response to management and Boron applications at Osceola 1 in small plots.

Table 4. The effect of Boron application and cut on alfalfa yield at Osceola small plots.

Table 5. Effects of management and Boron application on yield at Ionia in small plots.

Table 6. Effects of management and Boron application on ADF, NDF and CP concentrations at Ionia in small plots.

Potential Impact on Michigan Plant Agriculture/Industries:

This research will benefit Michigan alfalfa producers by generating agronomic information and economic data to provide comparisons between site-specific management and whole field management in alfalfa production.  Alfalfa fertilization with P and K may be accomplished more precisely as valuable site-specific yield data is collected through the use of yield monitors installed on hay mowers, near infra red aerial photography analysis, or satellite remote sensing imagery.  Knowledge of yield variation can help determine approximate nutrient removal from the alfalfa harvest within a field.  Knowing nutrient variability in fields allows for variable rate application of plant nutrients where they are needed rather than a one size fits all approach, which results in over and under fertilizing in the same field.  A case in point is represented at the KBS dairy farm where normal soil testing did not show the high P and K areas in the field, in fact, the normal soil testing procedure called for an additional 200 lbs/acre of K₂0 even though significant portions of the field already tested over 900 lbs/acre K.  This can also result in unfavorable environmental implications, especially when soil P is already at the maximum allowed under right to farm guidelines.  I