Spatial Analysis of Native and Sterile Pink Bollworm in the Arizona PBW
Eradication Program area, June 25 – July 22, 2006

By David Bartels, Ph.D., USDA,  APHIS,  PPQ, CPHST, Pest Detection,
Diagnostics & Management Laboratory, Moore Air Base, Bldg. 6414, 22675
N. Moorefield Rd., Edinburg, TX 78541-9398, and Michelle Walters, Ph.D.,
USDA APHIS, PPQ, CPHST, Decision Support and Pest Management Systems
Laboratory, 3645 E. Wier Avenue, Phoenix, AZ  85040

Introduction

The pink bollworm (PBW) moth (Pectinophora gossypiella (Saunders) feeds
almost exclusively on cotton plants (Gossypium spp.) and can cause
significant economic damage to the crop.  PBW is often the key pest of
cotton in the Southwestern United States and Northwestern Mexico. 
Currently, Arizona is in the process of eradication PBW from certain
areas using a combination of transgenic cotton, mating disruption with
pheromones, and sterile insect release.  The objective of the analysis
was to document the presence of native and sterile PBW populations in
the cotton production areas based on trap counts.

Materials and Methods

Native and sterile adult PBW were monitored in the Arizona PBW
Eradication Program cotton fields using standard Delta traps baited with
PBW sex pheromone.  Traps were checked weekly by program personnel and
numbers of male moths (native and sterile) were recorded.  For each
field, we were presented with the number of traps per field and the
total number of moths counted each week from June 25th – July 22nd,
2006.  The number of traps per field varied based on the size of the
field, and the type of cotton being grown.  The number of traps ranged
from 0 to 14.  A field may have no trap located in it because there is
one central trap that covers several small fields.  The average number
of moths per trap was calculated for each cotton field by dividing the
total capture by the number of traps.  To create a point for the trap
captures, the center of each cotton field containing a trap was
calculated using its geographic boundary.

To present the trapping data as a predicted surface of PBW numbers,
geostatistics was used to calculate a predicted value for areas between
the known values of each field.  Ordinary Kriging using a spherical
model was applied to trap counts for each week to develop a predictive
surface model surrounding the cotton fields.  Kriging is a technique for
the optimal interpolation of points across the spatial domain. Kriging
handles spatial autocorrelation and it is not sensitive to preferential
(uneven) sampling in specific areas, such as the distribution of cotton
field in the eradication program.  Kriging constructs a weighted moving
average equation that estimates the value of a spatially distributed
variable from adjacent values while considering the interdependence.
This equation minimizes the effect of the relatively high variance of
the sample values by including knowledge of the covariance between the
estimated point and other sample points within the range.  Kriging
results in a smoothing effect with high original values being
underestimated and low values being overestimated.  There is also less
variation than in the original points. Kriging is a best linear unbiased
estimator because it minimizes the variance of the estimation errors.

The Kriged surfaces were generated from a total of 3,476 center points
from the cotton fields.  We used a 2 kilometer range of influence from
the center point of the field, so that each field’s data is affected
only by other fields within approximately 1 kilometer of the outside
border of the field.  The 1 km limit reflects the perceived day to day
movement of PBW adults in green cotton fields and limits the undue
mathematical influence of a “hot” field on a large area.  In
addition to limiting the area of influence, we truncated trap values to
100 moths/field/week as this indicates a “hot” field biologically
and also, because a weighted average is used in Kriging, capping the
high values limits undue graphical influence of a single field.  A trap
with greater than 100 PBW moths also represents a unreliable count as
the efficiency of the trap declines quickly once the trap fills with
moths.

Results

The maps presented depict the native and sterile male PBW populations in
the Arizona PBW Eradication Program based solely on trap captures. 
Additional factors such as cotton type, soil conditions or crop
phenology are not included in the model.  The analysis indicates that
the sterile PBW adult population is clearly more abundant during each
week with many areas have populations of greater than 50 moths per trap.
 At the same time, the native populations is mainly around 1-5 moths per
trap with only 3 - 5 “hot spots” of numbers greater than 25 showing
up across the program area during the 4 weeks.  The sterile PBW
population is also more consistent from week to week and more widely
distributed than the native population.  The daily releases of sterile
PBW from the program aircraft have kept the sterile population levels
stable during the 4 weeks of analysis.

VOLUME 2

DOCUMENT 2

Page   PAGE  3  of   NUMPAGES  3 

