INTRODUCTION

Native ants play important roles in forested ecosystems by translocating and aeriating soil and contributing to litter decomposition and fragmentation.   Since its invasion in the 1930's, the red imported fire ant (Solenopsis invicta Buren) has displaced many native species and, consequently, has reduced native biodiversity, thereby altering natural ecosystems.  While many studies have been conducted on fire ant populations within different habitat patch types, none have focused on the interface between habitats.  As fragmentation of natural systems increases, these interfaces, or ecotones, will become increasingly important.  This study focuses on the distribution and interaction of fire ants and native ants along seven ecotones in a post oak savanna.

The goal of this study is to investigate the interaction of native ants and S. invicta along ecotones of a post oak savanna landscape.  The richness and abundance of native ants, and their correlation with S. invicta abundance will be compared between different ecotone types.

The objectives of this project are:
     1. To compare species diversity among seven ecotone types
     2. To identify any species that occur in only one ecotone
     3. To identify ecotones with a high species richness of ants
     4. To identify ecotones with a high abundance of S. invicta
     5. To identify any correlation between abundance of S. invicta populations and species
         richness of native ants



STUDY SITE

The Sawdust Ranch, owned by Kent Moore, is located on Texas Highway 1361 in Burleson County, TX. The 257-hectare (635-acre) ranch is located within the Post Oak Savanna Vegetation Zone and is currently being used for a deer hunting concession. Common plant species include post oak (Quercus stellata), blackjack oak (Q. marilandica), Smilax sp., and yaupon (Ilex vomitoria). Also abundant is introduced coastal Bermuda grass (Cynodon dactyolon). The official soil series is Burlewash, which is a moderately deep, well-drained, and slowly permeable upland soil.

 



METHODOLOGY

Spatial Database Development

Mr. Joe Pase (Texas Forest Service) flew the aerial photography mission of Sawdust Ranch in February, 1998. Twelve 9 X 9 in. color infrared (CIR) aerial photographs were acquired at 3,600 ft AGL. The CIR contact prints were created from positive transparencies and scanned at 600 dpi. The scanned images were georeferenced to the Snook DOQQ and resampled to 1m resolution. Images were histogram matched and mosaicked. Intergraph Image Analyst was used for all image processing. The blue X's on the image below represent the locations where aerial photographs were taken. 




Landscape Classification

A patch-corridor-matrix model was used to classify the landscape. In this classification system, seven patch types, a post oak woodland matrix, and a road system corridor were identified. The mosaic of habitat types represented an ideal site for studying habitat interfaces. The ecological classification coverage of Sawdust Ranch is shown below.



A 5m buffer of ecotones was created using ArcInfo software. This coverage was used to identify locations and measure lengths of ecotones on the ranch. Seven ecotone types were selected for study. Five transects were selected to sample ecotones with total lengths of less than 1,000m, 10 transects for ecotone lengths of 1,000-2,000m, and 20 transects for ecotones with lengths greater than 2,000m. This interval method assured a proportionate and representative sample of each ecotone on the landscape and simplified data collection and analysis. Ground-truthing the ecotone locations slightly reduced the transect numbers.



Site Location

Pitfall traps, aligned along transects, were used to sample the insects. A 5 pitfall trap design was used to sample across land ecotones and a 3 trap design was used for water to land ecotones. The study ranged 5m into the edge of each habitat type. Transects were laid perpendicular to the habitat interfaces. Holes were dug 2.5m apart and two 500ml cups were placed in each hole. The exposed cup and surrounding hole were backfilled with soil and left flushed with the ground while the trap was inactive. To activate pitfall traps, a cup with 140ml of a 50:50 water and propylene glycol solution replaced the soil-filled cup. A rodent exclusion device was placed on top of the traps. Traps were activated for a 7-day period, after which samples were collected and analyzed for ant species richness and abundance. Data was collected fall 1998 and early summer 1999.



Location of each pitfall transect was recorded using a Trimble Pathfinder II GPS unit. The different colored dots on the map below represent transects across different ecotone types.





RESULTS

Preliminary results indicate a statistically significant difference in S. invicta abundance between ecotone types. The wood-cultivated field ecotone had a significantly greater abundance of S. invicta than all other ecotones for the Fall 1998 sampling season. The grass-pond ecotone was significantly greater in S. invicta abundance than all but the cleared wood-grass ecotone for the Early Summer 1999 sampling season. When the data were averaged across sampling seasons, the grass-pond ecotone was significantly greater in S. invicta abundance than all but the wood-cultivated field ecotone. A difference in fire ant abundance between sampling seasons exists, but cannot yet be correlated to temporal or seasonal factors.




Eleven genera of native ants were found within the 7 ecotones studied: Camponotus, Crematogaster, Cyphomyrmex, S. (Diplorhoptrum), Dorymyrmex, Forelius, Hypoponera, Monomorium, Paratrechina, Pheidole, and Myrmecina. The most prevalent genera were Forelius, Hypoponera, and Pheidole. Solenopsis invicta abundance was significantly greater than native ant abundance for 6 of the 7 ecotones. Both the Fall 1998 and Early Summer 1999 sampling seasons showed similarities between the abundances of S. invicta and at least one of the native ant genera across the cleared wood-pond ecotone. The chart below displays the number of native ants and fire ants within each ecotone type, averaged over the two sampling seasons.





CONCLUSIONS

Preliminary analysis indicates a significant difference in S. invicta abundance across the seven ecotones. Grass-pond and wood-cultivated field ecotones had the greatest abundance of S. invicta. The abundance between ecotones also varied between the different sampling periods. Whether this difference is due to seasonal, temporal, or other changes will be analyzed in the future. Solenopsis invicta abundance was significantly greater than native ant abundance across 6 of the 7 ecotones. Wood-cultivated field is the only ecotone where S. invicta abundance is not significantly greater than all the other native ants found. Differences in ant generic diversity and abundance within each ecotone occurred between the two sampling seasons but has not yet been analyzed.

Future analyses will evaluate:
-native ant diversity and abundance within each ecotone
-distribution of native ants across the ecotones
-effects of S. invicta abundance on native ant diversity
-differences in ecotone structure that may attribute to greater native ant species richness
-differences in ecotone structure that may attribute to greater S. invicta abundance
-variability of data between sampling periods



ACKNOWLEDGEMENTS

Kent Moore - Owner of Sawdust Ranch
Dr. Douglas Wunneburger - GIS Technical Assistance
Texas Forest Service - Aerial Photography
U. S. Forest Service - Aerial Photography
Dr. Ben Wu - Geostatistician



Funding Provided by the Fire Ant
Research and Manangement Plan