Recipient Organization
UNIVERSITY OF FLORIDA
G022 MCCARTY HALL
GAINESVILLE,FL 32611
Performing Department
Agronomy
Non Technical Summary
The mission of this applied research program is to sustain management of aquatic systems in Florida by advancing technologies and intelligence towards optimizing resource inputs and improving performance outcomes. Experimental methods are deployed to study the performance of field applications, tactics and strategies in operational scenarios and whole lake systems. This uniquely creates direct coordination between the researcher and resource management practitioners for the adoption of new techniques and data analytics.The State of Florida has 1.25 million acres of public waters providing a broad range of services including, agricultural irrigation, hydropower, potable water, navigation, wildlife habitat, recreation and real estate. According to the Florida Fish and Wildlife Conservation Commission (FWC), 73% of these public waters are impacted by three category I invasive plant species: water lettuce (Pistia stratioides), water hyacinth (Eichornia crassipes) and hydrilla (Hydrilla verticillata). In fiscal year 2018, $17M in state and federal funding was dedicated to maintenance control of several category I invasive plant species. Of these, 80% of total area treated and 86% of the budget was targeted to floating plants (water lettuce and water hyacinth) and hydrilla. Since 2010, floating plants have made up only 7% of the herbicide budget, while 83% has been dedicated to hydrilla management. Failure to sustain maintenance strategies for each of these species would result in rapid lake impairment and expensive, long-term recovery.In January 2019 the Florida Fish and Wildlife Conservation Commission (FWC) paused their aquatic plant management (APM) program due to public perceptions of lake impairments resulting from the use of herbicides. A series of public "listening" sessions were conducted across the state. In response to these public testimonies, FWC identified a need for enhancements to their APM program: (i) Expand the creation of habitat management plans for individual lakes, (ii) Form a technical advisory group consisting of staff, partners and stakeholders, (iii) Improve timing of herbicide-based invasive aquatic plant removal treatment, (iv) Exploring ways to better integrate and increase the strategic use of mechanical aquatic plant harvesting, (v) Explore new methods and technologies to oversee plant treatments and (vi) Develop pilot projects to explore better integrated plant management tools (https://myfwc.com/news/all-news/aquatic-enhancements/). In each of the sessions there was a strong and consistent message to reduce herbicide use and deploy more mechanical harvesting. In response to this public demand, the FWC will initiate a pilot program to deploy mechanical harvester operations in order to reevaluate its contributions to control hydrilla compared with standard herbicide treatments as a complement to a more integrated management program.This project will monitor these operations to calibrate new standards for hydrilla management with herbicide and mechanical based tactics. The goal is to update management operations with technology upgrades producing high resolution spatio-temporal data that can generate new intelligence leading to better treatment precision, accounting and decisions.
Animal Health Component
80%
Research Effort Categories
Basic
(N/A)
Applied
80%
Developmental
20%
Goals / Objectives
ObjectivesThe purpose of this research is to understand the capabilities of mechanical harvest and herbicide treatment operations for better integration into the hydrilla maintenance program at Lake Tohopekaliga. This proposal has three objectives1. Calibrate hydroacoustic and airborne image remote sensing platforms for predicting hydrilla biomass and surface cover in managed and unmanaged lake systems.2. Conduct operations research on the mechanical harvest and herbicide treatment applications on hydrilla3. Monitor the terms of suppression of hydrilla to herbicide treatment and mechanical harvest applicationsIn accordance with the FWC plan to enhance the APM program, this project will coordinate with managers and vendors to adopt tracking technology that will generate new data and analyses calibrating operational efficacy. Further, we will develop a better measure of hydrilla growth response that will match the capacity of the harvest operations based on area and time of year. Finally, we will consider cost effectiveness of mechanical harvesting and standard herbicide maintenance treatments.
Project Methods
All experimental methods will be performed in conjunction with harvest and herbicide treatment operations to take place on Lake Tohopekaliga (Lake Toho). Lake Toho is a large (surface area 9,800 ha), shallow (mean depth 2.1 m), natural lake of the Kissimmee Chain in central Florida (Osceola County). This lake has an extensive management history with repercussions from non-point source nutrient pollution, organic sedimentation and aquatic plant invasion, leading to overall acceleration of lake succession (Hoyer et al. 2008, Hoyer and Canfield 1997).Hydroacoustic surveys will be performed from small watercraft to document percent biovolume of nuisance plants in the lake. Each survey boat will be mounted with a 20° beam transducer integrated with WAAS GPS (Lowrance Electronics; Tulsa, OK) set to log 200-kHz broadband signals at 10-15 pings s-1 with transverse boat speeds ≤ 3 m s-1. Acoustic and georeferenced data will be post-processed with BioBase® cloud-based software (Radomski and Holbrook 2015, Valley et al. 2015). Biobase algorithms will create spatial data layers of depth and aquatic plant height to calculate biovolume ratios of all georeferenced locations (Valley 2016). This transformed layer will be further interpolated with a kriging geostatistical algorithm to create a smooth biovolume estimation with pixel resolution equal to 0.2 transect spacing. Whole plots will be surveyed monthly for up to two years.Flight surveys will be performed with a DJI Inspire 2 vertical takeoff and land (VTOL) aircraft (DJI Technology Company; Shenzhen China) mounted with a X5S 20.8 MP CMOS camera with MFT 15mm/1.7 ASPH lens. This sensor offers sub-cm ground sampling distance (GSD) at an altitude of 45 m above the surface. A battery pair offers >30 min of sustained operational flight time. Mission planning will be performed with Pix4Dcapture software (Pix4D SA, Prilly, Switzerland) with standard 85% frontal and 70% side overlaps. All flights will be launched from watercraft in the vicinity of each region of interest to reduce ferry time and offer equivalent surface elevations translatable to flight plan altitudes. The RGB attributes for each pixel will be used to calculate Visible Atmospherically Resistant Index (VARI; Gitelson 2004) and Triangular Green Index (TGI; Hunt et al. 2011) as estimates of vegetation fraction and chlorophyll content.A vertical rake sampling method modified from Howell et al. (2017) will be deployed to measure SAV biomass at point locations. Samples will be cut at 50 cm below the water surface from a 1.0 m2. Fresh weights will be measured with a handheld spring scale immediately and further dried in a 70o C oven to complete dryness for final weight one week later. Samples will be further submitted for nutrient and carbohydrate analyses. This will be performed multiple times throughout the year to capture seasonal and phenological conditionsMultiple 1-ha observatories (n= 5-10) will be randomly established within three separate operational application areas, e.g., herbicide treatment, mechanical harvest or untreated infestation. Observatories will be surveyed monthly to calibrate hydroacoustic and airborne remote sensing platforms and calculate change detection over time. Hydroacoustic surveys will be conducted with transect spacings ranging from 5-20 m. Aerial images will be recorded over five different GSDs with an exponential range of 0.2-3.2. For both platforms, information gained is determined by the change in pixel value against the values established at the highest resolution (i.e., shortest spacing and GSD). A stratified random sampling design will be constructed from the quartile range of biovolume maps created one month prior for a total of 8 sampling points. Each sample will be sorted into individual stems. Stem lengths will be measured along with fresh weights and volumetric displacement to calculate individual stem density for empirical correlation of length to biomass. Monthly samples will be deducted from the previous month weights for a measure of relative growth rate.Non-linear regression and one-way ANOVAs will be performed with α=0.05 to model information gained as an effect of resolution for each of the remote sensing platforms, change detection biovolume and percent area cover over time. Pearson's correlation coefficients will be calculated to determine associations between hydoracoustic and airborne and biomass data. GPS and raster data will be compiled and analyzed using ArcGIS (v. 10.2.2; Esri; Redlands, CA) and Access and Excel (Microsoft Corp.; Redlands WA) with the XLSTAT add-in (v. 19.5, Addinsoft, New York, NY)For Objective 2, we will comprehensively measure space and time activities of a mechanical harvest operations and herbicide applications. For harvest operations this includes the harvester cutting and lifting hydrilla vegetation, a barge transporting the vegetation to shore and trucks hauling the material to the final disposal site. All operational equipment will be monitored with GPS tracking devices continuously recording locations on 10-sec intervals. This will display geo-reference data for all active and idle times of the operation. These devices will record the entire operations for up to 6 months. All GPS data will be post processed in ArcGIS (v. 10.2.2; Esri; Redlands, CA) to calculate several metrics:Total area harvested based on the travel distance and cutting swathHarvester speed as distance traveled over timeBarge travel distance and speedBarge time to load and unloadTruck travel distance and speedHerbicide applications will likewise be monitored with GPS tracking devices including the installment of a GPS + flowmeter-integrated logger that records where and how much herbicide was delivered to the designated areas. Comparable metrics to harvest operations include total treated area, applicator speed, while there are no barge and truck times to be estimated. The spatio-temporal elements of these operations can be transformed into basic economic consumptions (e.g., fuel, labor and herbicide) against the contractual arrangements (i.e., $ acre-1). These can be further utilized in a cost/distance analyses (ArcGIS v.10.2.2.; Esri; Redlands, CA) of Lake Toho for estimating optimal harvesting and herbicide options based on the shortest distances to shore access and/or disposal island and also the biovolume raster map translating into biomass and load count.Objective 3 will be a demonstration to compare the terms of suppression for mechanical harvesting and standard herbicide treatments. These will be established as un-replicated 50-ha plots on the southern portion of Lake Toho in monospecific "topped out" hydrilla. Treatments will be repeated in fall, winter and spring seasons. Plots will be spaced at least 500 m apart to prevent any herbicide overlap. Multiple machines will be deployed to complete the harvest operation within one week. The herbicide treatment (potassium endothall at 2ppm) will be applied within 1-2 weeks of the harvest.Monthly surveys will be conducted as described in objective 1 with 150 m transects (e.g., 10-20 m), starting one month prior to harvest and for six months after. Change detection will be measured from monthly percent area change and total volume growth over time (Valley 2016). Paired t-tests will be performed to measure the differences between time periods among the grid values for each treatment, separately, for determining when significant growth has occurred. The comparison between treatments of this un-replicated trial will utilize the relationship among the intraclass correlation coefficient (ICC), within and between-treatment variances and subsampling (i.e., grids) to carry out conservative tests of significance (Perret and Higgins 2006). Cost effectiveness will be determined by the variable costs of operations and the returns in suppression measured in days with a significant reduction in biovolume.