Datasets:
patrickbdevaney
/
mia9

Dataset card Data Studio Files
xet
 Community
1
text
stringlengths
0
6.44k
enrichment. The selection of these parameters finds support on a preliminary
assessment of FIU’s dataset indicating a positive correlation between both nutrients and
CHLa, especially TP, the recognized main limiting nutrient region-wide (Brand 1988;
90
Brand et al 1991; Boyer et al 1999; Fourqurean et al 1993; Szmant and Forrester 1996;
Fourqurean and Robblee 1999; Hoyer et al 2002; Boyer 2006; Boyer et al. 2009).
Threshold Analysis
In this report, we adopt the definition of ecological threshold of Andersen et al.
(2008) who stated …―an ecological threshold is the critical value of an environmental
driver for which small changes can produce an abrupt shift in ecosystem conditions,
where core ecosystem functions, structures and processes are essentially changed
between alternative states‖. Nutrient (TN, TP) concentration thresholds for each
segment were derived by identifying concentrations that were associated with sudden
and sustained increases (shifts) between CHLa alternative states. CHLa z-scored
cumulative sums were plotted along either TP or TN gradients, mimicking nutrient doseexperiment results. These graphs, constructed as described in Appendix 4, illustrate the
cumulative reaction of phytoplankton biomass to nutrient enrichment, highlight the main
threshold(s), and provide information to assess the potential health status of
phytoplankton communities in the water column.
Andersen et al. (2008) give an excellent description of driver:response scenarios
and the resulting regime shifts as shown in Fig 6.4. Panel (a) illustrates a regime shift
when the driver is linearly mediated to the ecosystem state response, and steps appear
only in the time series. Panel (b) shows a regime shift in ecosystem state after the driver
exceeds a threshold. The jump appears in the time series of the response. Panel (c)
shows a hysteresis loop linking the response to the environmental driver causing shifts
between two alternative states. The Zcusum charts along the lower row precisely
display the location of the threshold in every case.
There is a wide variety of patterns in Zcusum charts for South Florida waters
reflecting the complexity of these ecosystems as shown in a transect from SCI to ocean
(OFF) waters (Fig 6.5). In SCI CHLa seems to respond positively to nutrient
enrichment, although the pattern for TN is rather complex and the first important CHLa
reaction (shift) to increasing TN values occurs well below the median TN concentration.
This complexity reflects the dynamics of SCI, strongly affected by canal inputs, sudden
changes in salinity and nutrient concentrations, and perhaps different assemblages of
91
CHLa producers. For TP at SCI there is a CHLa increasing pattern (―V-shaped‖
Zcusum) and a threshold location way below the median TP concentration. Further
east, in SCM, both TN and TP seem to share the control on phytoplankton biomass
production, although the CHLa pattern is better developed for TN than for TP. Also
notice that for values just above the median TP concentration the relationship reverses
and CHLa declines. In SCO, more affected by oceanic exchange, TN and TP play
inverse roles on CHLa production with TN positively correlated and TP negatively
correlated to CHLa, and a potential threshold at relatively low TP concentration. The TP
threshold for SCI, SCM and SCO is the same, 0,004 mg/l TP. Finally, in oceanic waters
to the east and above the reef track (OFF) CHLa is positively correlated to TN and TP.
These relationships underscore both, the compartmentalization of Biscayne Bay waters
and the complexity of the exchange between relatively nutrient-rich freshwaters and
nutrient-poor oceanic waters.
Figure 6.4: Selected relations driver:response and their resulting regime shifts: (a)
smooth pressure–status relationships, (b) threshold-like state responses and (c)
bistable systems with hysteresis. Modified after Andersen et al. 2008.
Response Z-Cusum
Driver
Driver
Time
Response
Time
Response
Driver
Driver
Time
Response
Time
Response Z-Cusum
Driver
Response
Driver
Driver
Time
Response
Time
Response
Driver
Response Z-Cusum
Driver
a) Driver threshold b) Response threshold c) Driver-Response hysteresis
Modified after Andersen et al 2008
92
The relative location of the median nutrient concentration within the Zcusum
chart and its relation to the threshold position deserve special consideration. As shown
in Figure 6.5 median TN and median TP are above the CHLa thresholds, suggesting the
initiation of bloom conditions below median nutrient levels in the water column.
Furthermore, in all instances the median nutrient concentrations fall on the aboveaverage branch of the CHLa Zcusum chart (positive slope branch), indicating that
median nutrient concentrations are associated to above average CHLa production.
Finally, relationships for the overall region show that median TN and median TP are
above the CHLa thresholds almost everywhere. Those relationships suggest the
initiation of bloom conditions below median nutrient levels in the water column. Bloom
conditions simply means significant increase in biomass (Legendre 1990), abundance
(Lapointe 1999), or population size (Smyda 1997; Carstensen et al 2004; Sparrow
2007). In our case, the beginning of bloom conditions indicated by the threshold
correspond to the first sudden, significant and sustained CHLa increase.
Criteria
Although coastal and estuarine waters in South Florida have been affected by
human intervention and nutrient enrichment (Davis and Ogden 1994) most waterbodies
remain under oligotrophic-mesotrophic conditions, and most seem to meet with their
designated use. These characteristics have lead regulators to suggest numeric criteria
designed to maintain the current data distribution hoping to protect those uses in the
future (FDEP 2011). Our long-term trend analysis indicate that there are periods of
nutrient enrichment and associated water quality deterioration (i.e. algal blooms) which
may be linked to management practices, disturbances and/or climate variability. Hence,
statistics derived from the overall dataset, which includes these ―anomalous‖ periods,