Assessment of the effect of management activities like a drought salinity barrier and herbicide treatments on the spread of submersed and floating aquatic estuary macrophytes

Estuaries are complex ecosystems, being difficult to determine the way management actions affect them. This study quantitatively evaluated the spread of invasive submerged and floating aquatic macrophyte vegetation in Franks Tract of the Sacramento-San Joaquin Delta in response to two types of management actions, drought salinity barriers in years 2015, 2021 and 2022, and herbicide treatments in years 2004-2022. A Random Forest algorithm applied to airborne hyperspectral and satellite multispectral images generated maps of macrophyte cover in 2004-2022. Two methods assessed the effects of the management actions. The first one compared macrophyte cover trends between Franks Tract and reference sites across the Delta. The second one predicted macrophyte cover within Franks Tract using multiple environmental variables with a Long Short-Term Memory model. According to the comparison with Big Break reference site, submerged/floating macrophyte cover was +40.9 %/+8.6 % (2015), +57.8 %/+12.7 % (2021) and +75.7 %/+34.3 % (2022) with the barrier than predicted without it. Other reference sites rendered the same tendences in 2015 but mixed results for the other years. It is questionable if these sites behave as true references. Instead, the submerged/floating macrophyte cover was +8.0 %/+0.4 % (2015), +10.3 %/-20.6 % (2021) and -22.7 %/-9.3 % (2022) with the barrier than the Long Short-Term Memory model predicted without it. Additionally, the submerged/floating macrophyte cover with the herbicide treatments was -75.8 %/-72.4 % than modeled without it throughout 2004-2022. These models would improve quantitative assessments of management effects on macrophytes with a longer time series of data that includes more periods with and without drought barriers and herbicide use.