Introduction

The USDA-ARS Screwworm Research Program has found that in the tropics screwworm population densities are highest in ecotones between mature forest and open field. The USDA research unit in Panama is developing methods to identify these ecotonal habitats using LANDSAT TM and SPOT, satellite imagery. The ability to locate probable screwworm populations using remote-sensed data will increase the efficiency and effectiveness of USDA's biological control of screwworm.

SPOT XS Image   (click on name to view image*)   Landsat TM Image (click on name to view image*)

* These files are large and take alot of memory to view

SPOT, Landsat TM and DTED data were acquired of the study area. These data were evaluated for their usefulness in mapping screwworm habitats in the Panama Canal Zone, Panama using ERDAS Imagine 8.3 and PCI 6.0 software.

Field Data For Image Classification

Land use and vegetation data were collected in the field using a Magellan ProMark V GPS unit and a Magellan MBS-1 base station. For each location the vegetation was described by dominant species and percent vegetative cover at 1-5 meters, 5-10 meters, 10-20 meters and above. Ninety-six locations were visited in the Canal Zone Area. The GPS locations were used to collect training samples for image classification using ERDAS Imagine 8.2 seed pixel method.

Vegetation Map of SPOT and TM Image Analysis  (click on name to view image*)

* This file is large and takes alot of memory to view

Using ERDAS Imagine 8.3 a combination of supervised and unsupervised classification techniques were used on the imagery. The clouds and water were separated using cluster analysis. The vegetation types and land use were separated using the non-parametric decision rule. A number of band combinations with SPOT, LANDSAT TM and DTED were tried. The SPOT bands were considered the most accurate for discriminating vegetation types.  The Vegetation Map for the Canal Area is the combined results of the SPOT and TM classification.  The classification results from the Landsat imagery were used in areas of cloud cover on the SPOT image.    DTED data were used to separate the mangroves and highland forest types.  The red circles on the map indicate areas selected to sample for screwworm fly activity.

Field Data for Fly Activity

Panamanian technicians collected fly activity data in the different vegetation types identified on the imagery using a baited trap of rotting liver. They caught individual flies and took measurements of temperature and fly activity every 15 minutes. Data were collected from  5 to 7 days at each site.

To test whether vegetation type affect fly activity during the 3 seasons ( wet, transition and dry) we ran a two-way ANOVA on the data set of fly activity and vegetation type.  The area of highest fly activity is secondary forest between 20-30 m tall, vegetation code 6, and secondary forests more than 30 m tall, vegetation code 7.  In the Canal Area, vegetation code 7 tends to be where mature forests border open areas. The canopy in both vegetation codes 6 and 7 is often fairly open with moderate light penetration. The recorded temperature for these vegetation types were between 24 C and 30 C with a median of 26 C. The optimal temperature for screwworm fly activity is between 24 C and 26 C.  These results allowed us to easily map areas of high fly activity in the Canal Area.  We also found a high correlation between season and fly activity with the highest fly activity occurring during the transitions between the wet and dry seasons.

ANOVA Table