Degree

Doctor of Philosophy (PhD)

Department

School of Plant, Environmental, and Soil Sciences

Document Type

Dissertation

Abstract

Dense infestations of the free-floating aquatic fern giant salvinia (Salvinia molesta D.S. Mitchell) have been expanding across the southeastern United States for over two decades. Although herbicide programs have provided relief to giant salvinia infested areas, morphology and growth characteristics make management difficult. A series of mesocosm trials were conducted to explore alternative management strategies for giant salvinia. Evaluation of non-registered aquatic herbicides showed that rates of sulfometuron and metsulfuron controlled giant salvinia and provided a 98 to 100% biomass reduction. An outdoor mesocosm trial evaluated potential glyphosate resistance in giant salvinia in select water bodies throughout Louisiana. Differences between giant salvinia populations in response to glyphosate were not observed and a rate of 4.3 kg acid equivalent (a.e.) ha-1 was needed for 100% control. Outdoor mesocosm trials investigated herbicide efficacy against mature giant salvinia during the winter. In trial 1, flumioxazin, glyphosate, and glyphosate + diquat provided a 45, 55, and 45% biomass reduction, respectively, in comparison to the non-treated reference. In addition, January and February applications provided a greater reduction of biomass in comparison to December applications. In trial 2, giant salvinia biomass was reduced ≥ 99% for all treatments and timings in comparison to the non-treated reference. The effects of diquat, flumioxazin, glyphosate, and glyphosate + diquat over the top of bald cypress [Taxodium distichum (L.) Rich.] at one of three timings was evaluated. Delayed and abnormal leaf formation, reduced leaf length, irregular canopy formation, and no negative effects were observed among herbicide treated bald cypress. Biomass assessments and remotely sensed data, collected via satellite or drone platforms, were collected to determine if these technologies could be used to accurately predict herbicide efficacy. A linear regression analyzed changes in the near-infrared reflectance signatures of chemically treated plants and observed percent control of giant salvinia plotted against predicted control yielded an R2 = 0.843. These results indicate that remotely sensed data are a promising tool for monitoring large scale herbicide treatments for managing giant salvinia.

Date

4-2-2018

Committee Chair

Griffin, James

Included in

Weed Science Commons

Share

COinS