Estimation of submersed aquatic vegetation (SAV) biomass was evaluated using field data collected in 2017, and targeted analyses of three existing data sets: 1) Yin and Kreiling (2001), Drake et al. (2016), and 3) LTRM vegetation data (1998 – 2017). Two field studies were completed in 2017. The first targeted SAV biomass in raked plots and was conducted in collaboration with USFWS annual Lake Onalaska Vallisneria americana monitoring. In the second study, fresh weights of raked SAV were recorded at approximately 10% of LTRM Pools 4 and 8 2017 sampling sites.

System-scale restoration efforts within the Upper Mississippi River Restoration Program have included annual monitoring of submersed aquatic vegetation (SAV) since 1998 in four representative reaches spanning approximately 440 river km. We developed predictive models relating monitoring data (site-scale SAV abundance indices) to diver-harvested SAV biomass, used the models to back-estimate annual standing stock biomass between 1998 and 2018 and compared biomass estimates to previous abundance measures. Two morphologically distinct groups of SAV with differing sampling efficiencies were modeled and estimated separately: the first category included only Vallisneria americana which has long, unbranched leaves and dominates lotic environments, while the second category included 17 branched morphology species (e.g. Ceratophyllum demersum and Elodea canadensis) that dominate lentic environments. Two non-native species (Myriophyllum spicatum and Potamogeton crispus) were included in the branched category for model development but were estimated separately as a fraction of branched species biomass due to a small sample size and a lack of training data.

System-scale restoration efforts within the Upper Mississippi River Restoration Program have included annual monitoring of submersed aquatic vegetation (SAV) since 1998 in four representative reaches spanning approximately 440 river km. We developed predictive models relating monitoring data (site-scale SAV abundance indices) to diver-harvested SAV biomass, used the models to back-estimate annual standing stock biomass between 1998 and 2018 and compared biomass estimates to previous abundance measures. Two morphologically distinct groups of SAV with differing sampling efficiencies were modeled and estimated separately: the first category included only Vallisneria americana which has long, unbranched leaves and dominates lotic environments, while the second category included 17 branched morphology species (e.g. Ceratophyllum demersum and Elodea canadensis) that dominate lentic environments. Two non-native species (Myriophyllum spicatum and Potamogeton crispus) were included in the branched category for model development but were estimated separately as a fraction of branched species biomass due to a small sample size and a lack of training data.

The datasets are to accompany a manuscript describing the prediction of submersed aquatic vegetation presence and its potential vulnerability and recovery potential. The data and accompanying analysis scripts allow users to run the final random forests predictive model and reproduce the figures reported in the manuscript. Files from several data sources (aqa_2010_lvl3_pct_oute_joined_VEG_BARCODE.csv, eco_states_near_SAV.csv, ltrm_vegsrs_thru2019_GEOMORPHIC_METRICS_final.csv, vegetation_data.csv, and water_full.csv) were combined into a single .csv file (analysis_data_for_SAV_RandomForest.csv) used as the input for the random forest model. When intersecting points with geomorphic metrics some sites were moved slightly to ensure they were contained within aquatic areas (ltrm_veg_sites_moved.csv). Outputs from the random forest model are contained in the SAV_RandomForest_results.csv and SAV_RandomForest_results_testing_set.csv files.

The datasets are to accompany a manuscript describing the prediction of submersed aquatic vegetation presence and its potential vulnerability and recovery potential. The data and accompanying analysis scripts allow users to run the final random forests predictive model and reproduce the figures reported in the manuscript. Files from several data sources (aqa_2010_lvl3_pct_oute_joined_VEG_BARCODE.csv, eco_states_near_SAV.csv, ltrm_vegsrs_thru2019_GEOMORPHIC_METRICS_final.csv, vegetation_data.csv, and water_full.csv) were combined into a single .csv file (analysis_data_for_SAV_RandomForest.csv) used as the input for the random forest model. When intersecting points with geomorphic metrics some sites were moved slightly to ensure they were contained within aquatic areas (ltrm_veg_sites_moved.csv). Outputs from the random forest model are contained in the SAV_RandomForest_results.csv and SAV_RandomForest_results_testing_set.csv files.

This data set reports measurements of aquatic macrophyte biomass, phenology, leaf characteristics, and length and diameter of stems of both submerged and unsubmerged macrophytes. Data were collected from sites in the Monte Alegre Lake region on the eastern Amazon River floodplain in Para, Brazil. Ten field surveys were made at approximately monthly intervals from December 2003 to November 2004. There is one comma-delimited data file with this data set.

This data set reports measurements of aquatic macrophyte biomass, phenology, leaf characteristics, and length and diameter of stems of both submerged and unsubmerged macrophytes. Data were collected from sites in the Monte Alegre Lake region on the eastern Amazon River floodplain in Para, Brazil. Ten field surveys were made at approximately monthly intervals from December 2003 to November 2004. There is one comma-delimited data file with this data set.

The Long Term Resource Monitoring (LTRM) program employs a destructive harvest method for sampling aquatic vegetation whereby a rake is dragged ~1.5 m over the substrate and plant materials are retrieved. The density of each species of submersed aquatic vegetation (SAV), and of all species combined, are scored based on the amount of plant material collected on the teeth of each rake. Plant density (PD) scores are ordered and vary from 0 (no plants captured) to 5 (80-100% of rake teeth covered). The PD score of 1 has represented the vast majority of all non-zero values since 1998 and is associated with a wide range of biomass (e.g. <1g to 694g fresh weight in Pools 4 and 8 during the 2017 field season). However, small plant fragments account for a large proportion of individual species and combined species samples within this density score. This study evaluated a potential new “trace” PD score in the field in 2018 in Pools 4, 8 and 13 whereby trace was defined as any visible plant material up to 1/13th of a traditional PD score of 1.

The Long Term Resource Monitoring (LTRM) program employs a destructive harvest method for sampling aquatic vegetation whereby a rake is dragged ~1.5 m over the substrate and plant materials are retrieved. The density of each species of submersed aquatic vegetation (SAV), and of all species combined, are scored based on the amount of plant material collected on the teeth of each rake. Plant density (PD) scores are ordered and vary from 0 (no plants captured) to 5 (80-100% of rake teeth covered). The PD score of 1 has represented the vast majority of all non-zero values since 1998 and is associated with a wide range of biomass (e.g. <1g to 694g fresh weight in Pools 4 and 8 during the 2017 field season). However, small plant fragments account for a large proportion of individual species and combined species samples within this density score. This study evaluated a potential new “trace” PD score in the field in 2018 in Pools 4, 8 and 13 whereby trace was defined as any visible plant material up to 1/13th of a traditional PD score of 1.

This projects primary goal was to provide data on biomass of potential seed resources located within shallow water coastal areas within fresh to saline coastal waters of the northern Gulf of Mexico. The data set provides biomass of seeds, by species or lowest practical taxon from 2013, 2014 and 2015 across 384 randomly selected sites located in shallow water coastal areas. The data were collected between June and September of each year. This data set can be merged with a dataset which reports submerged aquatic vegetation and environmental data collected at the same time (La Peyre et al. 2017; https://doi.org/10.5066/F7GH9G44).