Implications of multi-scale sea level and climate variability for coastal resources: A case study for south Florida and Everglades National Park, USA
Trends and variability of water levels in the ENP
Water levels in the wetlands of ENP are influenced by sea levels, direct rainfall, and the amount of freshwater entering the Park across its northern boundary. Water management operations that control the amount of freshwater deliveries to ENP reflect regional strategies designed to balance the protection of fish and wildlife with agricultural and municipal water supplies and flood control. These strategies have changed over the past several decades, with distinct periods in water level regulations: the period 1961-1965 was marked by no freshwater releases during the construction of the control structures in the northern boundary of ENP (see Fig. 1) and was followed by a period (1965-1970) of relatively high discharges (25.98 ± 36.08 m3s-1). The subsequent period from 1971 to 1990 was marked by relatively low freshwater releases (15.25 ± 19.33 m3s-1), and a return during 1991-2008 to relatively high releases (33.33 ± 38.97 m3s-1). Prevailing rainfall conditions during these periods influenced the magnitude of these releases, although the degree to which rainfall and the freshwater releases were correlated varied over time because of changing water management strategies.
Water levels recorded at four long-term monitoring stations within ENP exhibit increasing trends approximately equal to or larger than the long-term trend at Key West (Fig. 3; Table 1). Figure 3 shows sea levels at Key West and water levels at the inland monitoring stations, as well as annual precipitation and freshwater releases measured at the northern boundary of the Park. The longest records (e.g. P33) show that the trends of water levels in ENP closely match the Key West sea level trend during the periods such as the early 1960s when overland flow across the northern Park boundary was largely blocked. Since the construction of gated control structures across the northern boundary and with the advent of managed releases (shown shaded in Fig. 3), the rate at which water levels have increased at the stations close to the northern boundary (P33 and P34) exceeds the rate of increase in sea level observed at Key West. This is evident in the inset scatterplot of the rate of water level change at the stations vs. the rate at Key West, where solid dots indicate the values in the period 1953-1965 prior to the construction of the release structures, and crosses indicate the post-1965 values. Note that the trends parallel the diagonal in the absence of managed releases, indicating that the sea level change signal at Key West is propagated inland. After the construction of the release structures, the trends at the stations are independent from the Key West trend, as shown by the crosses. For example, at P34 near the northern boundary of the Park, the 9.65 mm year-1 trend of increasing water levels is mainly due to increased freshwater releases after the 1960s. On the other hand, the relative effect of the freshwater releases is less obvious in water levels recorded at stations located further downstream, such as P35 and P37, where the trend has remained very close to the Key West trend over the period of record (2.56 and 2.27 mm year-1, respectively).
Cross-wavelet analyses show that interannual variability and correlation with the NINO3 and the NAO indices are also apparent in the coastal and inland ENP wetland stations (Fig. 4). The periodicities of joint high wavelet power are different between the Key West (Fig. 2) and the P35 analyses (Fig. 4), which is somewhat expected due to the fact that the P35 record spans the years 1954-2005; hence, the water levels are subjected to the additional effect of scheduled freshwater releases from the northern boundary and are not allowed to freely fluctuate in response only to natural variability. However, patterns similar to the ones observed in the Key West record (Fig. 2) emerge in the P35 record (Fig. 4), including the reversal of the NINO3 and Key West phases after the 1995 shift to the warm AMO. ENSO leads to increased wintertime precipitation in south Florida (Hagemeyer and Almeida 2002), and our results show inland water levels in ENP reflect this signal as a result of both direct rainfall and the timing and magnitude of structural releases across the northern boundary. ENSO contributes to the interannual variability in those stations further downstream and closer to the coast through its influence on both the Key West sea levels and the precipitation record, hence the high cross-wavelet power around the 4-years period in Fig. 4.
The observation of high power in the cross-wavelet spectrum of Key West sea level and the NAO Index, when the AMO changes phase, was also observed for the P35 case as indicated by the peaks in 1965 and 1995. We note a coincidence of high cross-spectral power in the periods 1965-1971 and 1985-1995, and accelerated rates of water level increase noted by the short trend lines in Fig. 3. This is attributed to the combined influence of ENSO and NAO on three of the primary factors that regulate water levels at this location, namely tidal inputs, precipitation, and managed freshwater releases at the Park's northern boundary.
|Fig. 3 The selected Park stations exhibit trends close to or higher than the Key West sea level trend. The barplot and shaded areas indicate precipitation and freshwater releases measured at the northern boundary of the Park. Short trend lines mark periods when water levels rise fast relative to the long-term trend. The inset figure shows the scatterplot of the rate of water level change at the stations versus the one at Key West: solid dots mark the values in years 1953-1965 (prior to the construction of the release structures), indicating that the trends parallel the diagonal in the absence of managed releases from the northern boundary of ENP. Crosses indicate the post-1965 values. For reasons of clarity, stations P36 and P38 are omitted [larger image]|
|Fig. 4 Cross-wavelet power spectrum between P35 water levels and a the NAO Index, and b the NINO3 Index. Inset arrows in the crosswavelet plot show the relative phasing of the two time series, with arrows pointing to the (left)right denoting (anti)correlation. Note the similarity in cross-wavelet power patterns (reversal of correlation sign, high power at the reversal of AMO phase etc) with the relationship between Key West levels and the NAO and NINO3 indices (compare with Fig. 2, see text for discussion) [larger image]|
|Table 1 Trends of water levels in the Park stations. Trends are calculated using the Sen's slope method (Mann 1945; Slack et al. 2003) from daily data over the period 1958-2008 for all stations, except for P36 (1968-2008). Elevation is relative to NAVD88|
|KW (Key West)||Marine||Florida keys||-||2.36|