Tidal data are the oldest and longest records available from the ocean, but much of the 19th and early 20th century tidal record (along with the weather data collected at tide stations) remains un-digitized and unused. In this project, novel ways of recovering, digitizing, vetting and analyzing 19th century tide data from Astoria, OR (1853-1876) and San Diego, CA (1853-1872) will be pioneered; these are two of the three oldest Pacific Ocean tidal records. The products of this work and related metadata recovery will be used to improve estimates of 19th century mean sea level (MSL) in the north-eastern Pacific, assess the secular change in tidal properties, and estimate long-term changes in storminess, for example, the climate induced, non-tidal variance that occurs due to wind, river flow, and other processes. This project will therefore address important contemporary scientific issues such as the long-term variation and acceleration in MSL, the long-term effects of climate change and local change, and the quality of old data.
The data will be improved (by removing errors) and the tabulated data that has already been recovered for Astoria and San Diego will be augmented by obtaining digital copies of selected marigrams (22 m long scrolls that each contain one month of pencil tide tracings) from the US National Archive. Metadata (leveling and datum records, time checks and weather records), will also be recovered and used to correct the raw data and to compare historic mean sea-level with modern measurements. One novel aspect is that recovered tide rolls will be digitized to 6 minute resolution using software, rather than the traditional slow and error prone hand tabulations. Further, innovative analysis techniques will be used to find and correct errors. Using the corrected data sets, MSL variation from San Diego, San Francisco, and Astoria will be assessed and compared. To the extent possible, corrections for barometric pressure, river flow, and isostatic rebound will be applied. The analysis will address whether MSL at San Francisco is representative of the US West Coast, and will, by comparison with the modern record, help constrain long-term trends in MSL variation. Secular shifts in the primary constituents (e.g., M2) and the shallow-water overtides (e.g., M4) will be assessed. Spectral analysis techniques and weather data collected at tide stations will be used to investigate secular changes to the non-tidal variance in the water level signal for San Francisco. Because overtides are produced by bathymetry, river flow, stratification, and other local processes, assessing their change over time can provide insight into locally forced changes to the functioning of estuaries. To the extent possible, the effects of local changes, climate, and river forcing will be separated.
Broader Impacts: A better understanding of changes to MSL and storminess helps address important contemporary scientific questions with policy implications. Because the proposal is a proof of concept, the methods to be developed here can in the future be used to recover hundreds of station-years of the 1850-1950 tide data that are currently stored in analog form (usually as marigrams) in archives. The PI is an early career scientist.
Tide data are the oldest oceanographic data available and measure the effects of storms, floods, tides, and sea-level rise in the coastal zone. Despite their importance, most nineteenth century records from North America have long been forgotten in archives, in many cases unused in over a century. During this project we produced over 50,000 photographs of archival data stretching back to 1835. We then digitized approximately 60 station years of nineteenth century west-coast tide data, and 80 years of New York harbor tide data back to 1844. Approximately 10 years of tide data from Astoria, OR, were digitized to 1 minute frequency via processing of the original pencil trace on marigrams, or tide rolls (Figure "Marigram Example"). Important results from the project are: 1. Storm tides and storm surge in New York harbor have greatly increased since 1844. The once-in-10 year storm tide (the measured water level minus sea-level) has increased by nearly 0.3m since the nineteenth century. Including local sea-level rise, the 10 year storm-tide has increased by more than 0.7m, and the risk of exceeding the 1.75m tall Manhattan sea-wall has increased by 20x (see figure labeled "Changing Storm Tides in New York Harbor"). These results were published in Geophysical Research Letters and were widely reported in the media (www.google.com/#q=manhattan+seawall&tbm=nws ). 2. The return period of hurricane Sandy level flooding is estimated to be 200-400 years, based on analysis of hurricanes from 1844-2013, and including the large hurricane from 1821 (Talke et al., 2014 in prep). Without the 1821 event, return periods are greater than 1000 years. Since a hurricane in 1788 was nearly as large as 1821 and Sandy, the smaller return period is considered more plausible. See Fig. "New York Hurricane Return Period". 3. Mean sea-level in Astoria shows no trend in either the nineteenth century (1853-1876) or in the modern period (1925-2013). These results confirm that vertical land-movement in the Columbia River Estuary is keeping pace with global sea-level rise. Sea level in San Diego has steadily increased since 1853 (Fig. "West Coast Sea Level, 1853-present"). 4. Tidal properties in the Columbia River, Hudson River, and Ems estuaries show evidence of continual perturbation since the 19th century. Increased channel depth, wetland reclamation, and altered river flow have likely caused changed tidal properties (Fig. "Tidal changes, 1800 to present". We hypothesize that storm surge properties have also been affected by local changes since the 1800s. 5. Annual River flow into the San Francisco Bay area has decreased by 30% since the 19th century, based on estimates from tide data (Moftakhari et al., 2013). Large snow-melt driven flows in spring-time have virtually disappeared (Fig. "River Flow estimates from Tides"). 6. Tide data show that the Dec. 1861 and Jan. 1862 floods in OR and CA, respectively, were approximately 25% larger than any subsequent winter flood (Moftakhari et al., 2013; Talke et al., 2014 in prep; Fig "River Flow estimates from Tides"). 7. The storm-tide hazard in New York fluctuates with large-scale atmospheric fluctuations in the North Atlantic. Historically, the risk of large storm-tide events is smaller during some periods (1920s) than in other periods (1950-1970). Combined with other factors such as sea-level rise, the storm-tide risk in New York is continually changing (see Figure "Changing Storm Tides in New York Harbor"). In summary, historic 19th century water level data provide new insights into mean sea-level trends, tidal dynamics, storm tides, and river flow over secular (century scale) time periods. Understanding such trends may help scientists better predict the possible consequences of climate change and continued development of harbors. If storm-tide risk has increased due to local processes, there may also be local solutions in some cases. In the future, the recovered tide data can help address important scientific controversies, such as whether sea-level acceleration is occurring and whether climate change is increasing storm magnitudes.
