text stringlengths 0 6.44k |
|---|
is present in our individual SoFlo basins. This fact called for a holistic approach to basin |
segmentation to account for variability not only dictated by a given nutrient |
concentration level, but by the combination of imposed conditions such as: nutrients, |
water clarity, climate, weather, extreme events, geomorphology, circulation and |
exchange, and management. As done for Biscayne Bay, basin segmentation was |
accomplished with an objective classification of station sites combining Principal |
Component Analysis (PCA) and Hierarchical Clustering methods in tandem. Selected |
biogeochemical variables (8 to 13) for each of the five basins (Florida Bay, Florida Keys, |
Whitewater Bay-10,000 Islands, Gulf Shelf and Pine Island-Rookery Bay) were used for |
PCA. Statistics (mean, standard deviation, median and median absolute deviation) of |
retained scores from PCA were input into Hierarchical clustering routines. The rationale |
behind the selection of these statistics was to account not only for level of individual |
parameter but also for their variability. Then, progressive subdivisions were obtained |
varying the statistical distance among groups in the cluster tree. Selection of the final |
number of segments was subjective and knowledge-based. Spatial extension and |
pattern, geomorphology, water circulation and benthic ecosystem distribution played a |
major role on the decision. In summary, we statistically characterized and subdivided |
86 |
waters in each basin into what we considered biogeochemically and spatially coherent |
segments. |
We identified 40 water-types for SoFlo coastal and estuarine waters that extend |
from Biscayne Bay in the east to Dry Tortugas in the southwest and to Pine Island |
Sound in the northwest (Table 6.1, Fig 6.3). Segment maps were generated placing the |
separating lines approximately midway in between clustered stations, or followed the |
FATHOM Model subdivision (Cosby et al. 2005), followed previous subdivisions or were |
arbitrarily drawn following general geomorphologic patterns. In summary, border lines |
are site-specific and not the result of systematic spatial statistical analysis. |
Non-parametric Mann-Whitney and Kruskal-Wallis tests were used to compare |
biogeochemical variables among segments, and box-and-whisker plots to summarize |
descriptive statistics. The former analyses highlighted statistically significant differences |
among segments, and the later underscored anomalous and probably impacted stations |
and segments. Sustained high concentration levels would suggest sustained impacted |
and perhaps impaired conditions – not meeting designated use type. Anomalous |
stations isolated from permanent sources of human disturbance (urban areas, canal |
mouths) were interpreted as responding to site-specific natural behavior. On the other |
hand, eventual occurrence of high nutrient levels would indicate anomalous conditions |
responding to short-lived disturbances. |
Besides those precipitation changes occurring in the early-to-mid nineties from |
lower to higher precipitation rates which lead to region-wide declines in water nutrients |
and chlorophyll a concentrations (Briceño and Boyer 2010), we compared nutrients |
time-series with records of storms and hurricanes which affected South Florida during |
the period 1989-2008. Separate statistics were calculated for before and after identified |
events to test the nature, duration and magnitude of the impact. After identifying that |
human and hurricane impacts combined to strongly disturb conditions in northeastern |
Florida Bay-Southwestern Biscayne Bay since 2005, we discarded data obtained from |
that area (MBS) during the disturbed period (2005-2008) for statistical calculations. The |
dataset resulting from this selective process defined the Period of Calculation (POC) |
dataset to be used for the remaining calculations. |
87 |
Table 6.1: Aggregation of SoFlo basins into 40 segments resulting from PC/Cluster |
Figure 6.3: Biogeochemical segmentation of South Florida’s estuarine and coastal |
waters. |
Region No. Segment Sub-basin Region No. Segment Sub-basin |
1 MAR Marquesas 1 CS Card Sound |
2 BKB Back Bay 2 NCI North central Inshore |
FLORIDA 3 BKS Back Shelf 3 NCO North Central Outer-Bay |
KEYS 4 LK Lower Keys BISCAYNE 4 NNB Northern North Bay |
(FK) 5 MK Middle Keys BAY 5 SCI South Central Inshore |
6 UK Upper Keys (BB) 6 SCM South Central Mid-Bay |
7 Offshore Offshore 7 SCO South Central Outer-Bay |
1 CFB Central Florida Bay 8 SNB Southern North Bay |
FLORIDA BAY 2 ECFB East-Central Florida Bay 9 MBS Manatee-Barnes Sound |
(FB) 3 NFB North Florida Bay 1 CI Collier Inshore |
4 CL Coastal Lakes 2 EB Estero Bay |
5 SFB South Florida Bay PINE ISLAND 3 MARC Marco Island |
6 WFB West Florida Bay ROOKERY BAY 4 NPL Naples Bay |
1 BLK Black River (PIRB) 5 PINE Pine Island Sound |
WHITEWATER 2 CTZ Coastal Transition Zone 6 SCB San Carlos Bay |
BAY 3 GI Gulf Islands 7 COCO Cocohatchee |
(WWB-TTI) 4 IWW Internal Waterways SHELF 1 IGS Inner Gulf Shelf |
5 MR Mangrove Rivers (SHELF) 2 MGS Midddle Gulf Shelf |
6 PD Ponce de Leon 3 OGS Outer Gulf Shelf |
7 SRM Shark River Mouth |
8 WWB Whitewter Bay |
OFF |
88 |
NUMERIC NUTRIENT CRITERIA |
As a contribution to the regulatory process under way for the derivation of |
numeric criteria by the USEPA and FDEP, we have: (1) Sub-divided the six major South |
Florida basins into forty spatially-coherent water-types (segments). This classification |
has been adopted by both agencies with minor modifications; and (2) we have derived a |
series of protective nutrient thresholds to be considered in the derivation of protective |
numeric nutrient criteria. |
The USEPA recommends three types of approaches for setting numeric nutrient |
criteria (USEPA 2001, 2010): (1) reference condition approaches; (2) stressor-response |
analysis; and (3) mechanistic modeling. The reference conditions approach entails the |
existence of undisturbed segments of reference, something not to be found in South |
Florida after over a century of continuous human intervention. A suggested alternative is |
to select the less impacted station or segment complying with its designated use as |
reference, but given the recognized site-specific response to nutrient variability that |
characterizes estuaries, there is no guarantee that the extrapolations would be correct. |
The stressor-response approach requires cause-effect data (i.e. nutrient-dose |
experiments), and that information is mostly absent or very scarce for relevant South |
Florida biotic components such as phytoplankton biomass (Brand 1986; Brand et al. |
1991) or ambiguous for major benthic primary producers (seagrass, epiphytes, |
macroalgae, and benthic microalgae) (Armitage et al. 2005; Ferdie and Fourqurean |
2004; Herbert and Fourqurean 2008). Findings from a handful of studies in Florida Bay |
indicate that N addition had little effect on any benthic primary producers throughout the |
bay, and P-induced alterations of community structure were not uniform and did not |
always agree with expected patterns of P-limitation (Armitage et al. 2005). Nutrient-dose |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.