A paleo-proxy-based reconstruction of still water level in the San Francisco Bay during 1500 - 2000 CE

The dataset contains the following files: 

AWT_hisPCs_1880_2017.csv : Three leading annual principal components (APCs) of sea surface temperature anomalies (obtained from ERSSTv4) in the Hovmoller space during 1880 to 2017. 

APC*_1000sims_500yr.csv : Reconstructions of three APCs using paleo-proxy-based Nino3.4 reconstructions from 1500 to 2000. Each file contains 1000 simulations. 

TESLA_SF_reconstructed_SWLs_Sims_*_*_wSS.mat : there are four files for total 1000 simulations of reconstructed still water levels (SWLs) at San Francisco Bay. They can be accessed using MATLAB. Each of the files contains the following data:

time_fin : time in matlab format, 1500 - 2000 hourly. Time series starts from June 1st 1500, to account for wave climate. 

tide_raw_fin : astronomical tide prediction from 1500 - 2000, at hourly resolution. 

tide_raw_slr_fin : adding ~2.2mm/yr of SLR on to tide_raw_fin starting in 1900. 

tide_fin : correcting tide_raw_slr_sin with an ~0.09m offset to also account for the mismatch between MSL from ~1900- 2000 and the time before SLR is calculated. 

mmsl_sim_hourly : monthly mean sea level, anomalies plus seasonality, at hourly resolution. 

SWL : still water level at hourly resolution (tide_fin + mmsl_sim_hourly)

SS_fin : hourly storm surge time series. 

DSWL : SWL + SS_fin. 

MMSLA_100yrsRTs_30sw.RDS : Results of quasi-non-stationary flood risk analysis using annual maxima series (AMS) of MMSLA for a return period of 100yrs. The data set contains the following: 

eva: 1000 x 3 array contains GEV estimation of return levels for each of 1000 simulation. Column 2 contains the estimated return levels, columns 1 and 3 contains 2.5 percentiles and 97.5 percentile values around the estimates. 

swa : contains the results of GEV calculated over a 30 years sliding window. This is a list of 6 components as below: 

RTsim : An array of return level estimated for j-th window and i-th simulation 

RTobs: Return level estimates for last 100years (historical period)

ci.lower.delta and ci.upper.delta : these denote the 95% interval of sampling variability in 100yr events using the Delta method. 

window.size : size of the sliding window. 

sliding.yrs : starting year of each sliding window. 

conf : alpha or significance level (0.05 in this case).

return.period : return period (100 years in this case). 

 method: character string denoting which parameter estimation method is used ("MLE"). 

type: character string denoting which extreme value distribution is used *("GEV" or generalized extreme value distribution in this case). 

ams.ts : annual maxima series (input) values are saved here in the output. Can be used to reproduce the analysis. 

cur.ts : NULL. If ams.ts is not supplied, ams.ts is calculated using the input values of cur.ts. Not stored in the output if ams.ts input values are supplied. 

MMSLA_100yrsRTs_100sw.RDS: Results of quasi-non-stationary flood risk analysis using annual maxima series (AMS) of MMSLA for a return period of 30yrs, with a sliding window of 100 years. The components of this .rds are same that that for 100yrs sliding window.  

ams_dswl_1_1000_sim_wSS.rds : Annual maxima series of SWL (MMSL + tide + Storm surge) for 1000 simulations as simulated using TESLA. 

ams_dswl_1_1000_sim_wSS_no_slr.rds :  Annual maxima series of SWL (MMSL + tide + Storm surge) for 1000 simulations after removing SLR trend from 1900 onwards.  

ams_dswl_1_1000_sim_wSS_all_slr.rds :  Annual maxima series of SWL (MMSL + tide + Storm surge) for 1000 simulations after adding a total  0.2 m SLR over 1500 - 2000 CE to superimpose SLR on to full extent of natural variability.  All .RDS files are to be accessed using R. 

For further details, please contact Sudarshana Mukhopadhyay ([e-mail])