2.4. Comparison of Post-Harvey Sediment and Metal Concentrations to Historical Data
The GBES was divided into three independent segments as shown in Figure S2 in the SI to account for the various hydrodynamic regimes; a riverine segment, a mid-bay segment, and a bottom-bay/outlet segment. All segments receive waters from Houston watersheds while Trinity Bay, East Bay, and West Bay (Figure S2 in the SI) were excluded from the analysis because they are affected by the Trinity River, the Intracoastal Waterway and Chocolate Bayou, respectively.
Sediment Grain Size and Transport Associated with Extreme Events . Data for several extreme events in the Houston area were used to elucidate observations related to changes in sediment and metals regimes after extreme events; the events included: Tropical Storm Allison (Allison, 2001), Hurricane Ike (Ike, 2008), Memorial Day Flood (MD, 2015), and Tax Day Flood (TD, 2016), and Hurricane Harvey (Harvey, 2017). Due to a lack of samples immediately before and after some of the events, samples collected one year before and after the events were used. The MD and TD floods occurred within a year of each other, thus for the purposes of the analysis in the paper, they were combined using a time period one year before MD to one year after TD flood. Similarly, the TD flood and Hurricane Harvey were combined, the data that was available between these two events was used as “after event” for TD flood and “before event” for Hurricane Harvey.
For each segment in the GBES shown in Figure S2 in the SI, average values of all existing data within the segment and before and after each event was used. The percent change in metal concentrations in sediment and changes in each grain type were calculated using:
\begin{equation} \text{Percent\ changes}\left(\%\right)=\frac{\left(\ value\ prior\ to\ the\ event-\ value\ after\ the\ event\right)\ }{\text{value\ prior\ to\ the\ event}}\nonumber \\ \end{equation}
The Pearson correlation test was performed on the percent change in metal concentrations and the percent change in each grain type. Arsenic and Cadmium were excluded from the correlation analysis because 46% and 58% of historical sediment samples showed values below the detection limit for arsenic and cadmium, respectively.
Additionally, and because of the unique volume of transported sediment during Harvey, additional analyses were undertaken for sediment data collected after Harvey that included visual inspection, repeated measures ANOVA (RM-ANOVA), and paired t-test statistical analyses using IBM ® SPSS ® (IBM, 2017) to place the data in context relative to historical patterns.