Organic geochemical proxies (OGPs: δ13C, δ15N and C/N) preserved in coastal lake sediments appear to be more sensitive indicators of past storm events than traditional proxies used in paleotempestology. However, the method has not been tested in many lakes and with modern data. In this research, we measured the δ13C, δ15N and C/N values of suspended particulate organic matter (POM), along with salinities and stable isotopic compositions of water (δ18O and δ2H) from two coastal lakes in north and northwest Florida on seasonal or shorter time scales throughout a 3-year period (from May 2016 to October 2019). The time-series geochemical data not only show that geochemical properties of these lakes varied seasonally, reflecting variations in lake biological and environmental conditions, but also displayed unique variation patterns in response to large storms that caused either seawater (SW) flooding or freshwater (FW) flooding of the lakes. Our data show that SW flooding led to higher δ13C and δ15N values, with either lower or no change in the C/N ratios of the POM, generally consistent with a previously proposed conceptual model for detecting SW flooding events. The data also show that FW flooding reduces δ15N and increases the C/N values of POM, and lowers the salinity, δ18O and δ2H of lake water. These modern time-series data demonstrate the feasibility of detecting past storm events that were large enough to cause either SW or FW flooding from analysis of OGPs preserved in coastal lake sediments. Applying this understanding, we reconstructed a high-resolution proxy record of storm history from a coastal lake (Mullet Pond) in North Florida over the last 4500 years. The OGP-based storm record suggests 30 flooding events over the last 166 years, which can be matched (within the dating uncertainty) with almost all of the historic hurricanes that are known to have passed within 150 km of the Mullet Pond site in north Florida. This further confirms that OGPs preserved in coastal lake sediments may be used as reliable recorders of past storm activity. The OGP record reveals three active periods with storm frequencies near this site ≥ 15 storms/century based on both SW and FW flooding events, and two particularly quiet periods with <1 or no storm event detected during the past 4500 years. The OGP-based reconstruction of storm activities suggests that storm activity peaked between 1120 Cal yr BP and 1380 Cal yr BP with storm frequencies ≥ 21 storms/century, and the peak of this active period is offset by ~300 years (due to radiocarbon age uncertainty) from the Medieval Warm Period (MWP: ~900-1300 C.E.). The active storm interval from ~990-820 Cal yr BP (960-1130 C.E.) coincides with MWP. The peak active period between 1380 and 1120 Cal yr BP in the OGP record exceeds the storm frequency in the post-1850 era, consistent with the statistical model predictions of past tropical cyclone activity based on instrumental and other proxy-reconstructed climate indices. We have also applied the OGP approach to a similar coastal lake the Cedar Key area in the northeastern Gulf Coast of Florida and developed a high-resolution proxy record of storm history over the last 2300 years. The OGP record reveals five active periods based on both SW and FW flooding events, with storm frequencies near this site are greater than the mean storm frequency (2.8 storms/century), and three quiet periods with <1 storm/century have been detected. The OGP-based storm reconstruction suggests that storm activity peaked between -60 Cal yr BP (1890 C.E.) and 75 Cal yr BP, which is consistent with storm reconstruction from southwestern Florida and Caribbean sites, with relatively warm sea surface temperatures (SSTs) in the Gulf of Mexico. The two other active storm intervals (between 290-480 Cal yr BP and 520-640 Cal yr BP, the later one roughly corresponds to MWP) at Cedar Key site between 1310 and 1660 C.E. show frequent hurricane activities, likely due to warm SSTs, that are also consistent with other proxy records from GOM sites, the North American east coast, and the Bahamas. The results from both the study sites are generally comparable to those of other studies using different storm proxies. These results support the concept that long-term hurricane activity is cyclical and likely driven by variations in large-scale climate and oceanographic conditions.
