The Tides they are a Changin (With Apologies to Bob Dylan) David A. Jay Department of Civil & Environmental Engineering Portland State University Portland, OR 97207 USA Thanks to Richard Ray, Rich Pawlowicz,, Keith Leffler and Ed Zaron Research supported by the National Science Foundation, US Army Engineers E (Portland District), and NOAA-Fisheries
Why do we Care About Evolving Tides? Basic science evolving tidal amplitudes: Change mixing over shelves, in estuaries, maybe in open ocean Mixing affects currents, salinity, productivity and DO levels Illuminate the processes interacting with the tides May serve as one more indicator of climate change Define historical changes in estuarine and coastal environments and processes the longest records we have Real-world applications changes in coastal tides: Are important to navigation Affect location and character of inter-tidal tidal habitats, including those of commercial shellfish species Alter juvenile salmon habitats Influence extremes of coastal inundation coastal flooding due to storms, waves and high tides
Topics for Today Methods Changes in tides throughout the Eastern Pacific Speculations about causes Some interesting examples San Francisco Bay changing continental shelf processes? Seattle small changes hard to pick out of environmental variability Astoria very strong anthropogenic changes
Methods How to extract long-term trends in tides constituents
Methods Extract amplitude and phases of largest diurnal (K 1 ) and semidiurnal (M 2 ) from data and tidal potential by: Annual harmonic analyses, or Complex demodulation (3-yr convolution filters) Remove 18.6 yr cycles from amplitude and phase time series: Use admittance (complex ratio of tide to potential): Calculate complex admittance at each frequency Fit trend to result and multiple by mean admittance,, or Fit 18.6 yr cycle + trend to amplitude and phase Two approaches to tidal potential: Godin tabulation (embedded in t_tide) NASA code based on Cartwright and Eden (1973) Did not focus on S 2, to avoid radiational effects S 2 changes are larger at most stations where it was analyzed Analyzed ~80 station in Eastern Pacific; final results based on 34 coastal stations with records >43 yrs Red = methods used to produce final results
Methods (Continued) Final Procedure: Hourly data Hourly tidal potential Overlapping 3yr convolution filters for MSL rise, M 2 and K 1 (S 2 at some stations) Annual amplitudes and phases for M 2 and K 1 (from data) Annual amplitudes & phases for M 2 and K 1 (from potential) Complex admittance[ω] =data/potential Admittance trends Amplitude and phase trends MSL rise rate Normalized trend = trend/amplitude Total D 1 and D 2 rates D 1 =diurnal D 2 =semidiurnal Total Rise of HW levels
Long-term Changes in Eastern Pacific Ocean Tides
Eastern Pacific Amphidromes Amphidromes are the rotary structures around which tides propagate Tidal waves rotate around amphidromic points because of the rotation of the earth (via the Coriolis force) Each tidal wave rotates about an amphidromic point with zero amplitude There are three major K1K and M 2 amphidromes in the Eastern Pacific affect coasts of N. and S. America The lines on the plot at right are lines of constant phase (time of high water), emanating from the amphidromic points K 1 From David Pugh s book M 2
Station Locations All coastal stations with >43yrs LOR Excludes Hawaii and other islands; tides are growing there, also, especially M 2 There is a shortage of long records south of Acapulco Improve method to use shorter records!
Amplitudes and Rates This is all stations >43 yrs Patterns for both M 2 and K 1 : NE Pacific tides are growing M 2 is decreasing in Gulf of Panama S Pacific tides may be growing Also, lots of local variability South Pacific South Pacific Gulf of Panama M 2 K 1 NE Pacific amphidrome Gulf of Panama NE Pacific amphidrome
Relative Amplitudes Makes patterns clearer With the exception of two stations in Mexico with noisy records, both M 2 and K 1 are increasing ~2.2%/100yrs north of 19ºN It is a surprise that M 2 and K 1 are acting in the same way, more or less Why??? Century -1 0.12 0.08 0.04 0-0.04 K 1 M 2 M 2 and K 1 Relative Rates 19ºN 2.2% San Francisco Astoria Queen Charlotte City 0 5 10 15 20 25 30 35 Station Number Seattle Victoria Vancouver
Total Increase in HW Levels and MSL Rise Local MSL is highly variable, due to tectonics (not corrected for r here) MSL rise dominates south of San Francisco Increase in tidal amplitude is significant in NE Pacific where local l MSL rise is small (on average) mm Century -1 400 300 200 100 0-100 Tidal,Tidal +MSL, and GSL Rates D 1 + D 2 +local MSL GSL Rate D 1 + D 2-200 0 5 10 15 20 25 30 35 Station Number
Possible Causes of Large-Scale Tidal Evolution The open ocean is a large, slowly changing environment These changes are very rapid by oceanic standards The total tidal energy loss is almost constant Set by astronomy Changes VERY slowly Observed Changes are 100-10,000 10,000 times faster than changes in astronomical forcing So what can cause amphidromic-scale changes in tides, without changing the total tidal energy loss?: Moving amphidromic points (but why?) Changes in barotropic-to to-baroclinic energy conversion (BBEC) This is the loss of tidal energy to internal motions, including internal tides Similar changes are occurring in the North Atlantic
Causes of Large-Scale Tidal Evolution (More) We have hypotheses, not definite answers: Mechanisms could be large-scale (amphidromic( to basin sized), or Smaller-scale, but affecting larger areas Possible large-scale mechanisms: Changes in background currents at key locations (e.g., a shift in i the West-Wind Wind drift?) Changes in upper-ocean density stratification could alter BBEC Smaller-scale processes with amphidromic-scale impacts: Changes in stratification Around mid-ocean ridges this is happening in Hawaii Over continental shelves (e.g., increased fresher water input to Alaska coastal current) Changes in critical slope at shelf break, which would change internal tides Analysis is complicated by small-scale, scale, local changes; e.g., harbor modification and shelf internal tides
Some Pacific Northwest Examples San Francisco, Seattle, Astoria
m Long-term Changes in SF Bay: K 1 Record extends from 1899-2008 MSL is rising at 0.2m/100yrs K 1 tide is larger & earlier Amplitude trend is small, ~3±2.1 mm/100 yrs K 1 residual is small, but shows some 18.6yr signal that is not removed by any method 1.1 1.05 1 0.95 0.9 0.85 0.8 0.75 Mean Sealevel and trend (MSL) 1920 1940 1960 1980 2000 0.3 0.29 0.28 0.27 0.26 0.25 deg 0.45 0.425 0.4 0.375 0.35 0.325 0.3 50 45 40 35 30 25 K 1 admittance amplitude+trend 1920 1940 1960 1980 2000 K 1 amplitude: trend +18.6 yr cycle 1920 1940 1960 1980 2000 K 1 phase 1920 1940 1960 1980 2000
Long-term Changes in SF Bay: M 2 M 2 tide is getting larger and earlier Strong amplitude increase, 40± 5mm/100 yrs Phase: -6.6 ±2.4º/100 yrs (~13 min) S 2 is increasing at 15 ±1.2 mm/100yrs, proportionally faster than M 2 Could be local dredging or Atmospheric Not clear why SF Bay M 2 is noisier than K 1 even though K 1 is much smaller: M 2 more affected by offshore internal tides???? Large residual in recent extreme ENSO years, 1982-83, 83, 1995-1999 1999 0.42 0.4 0.38 0.36 0.34 0.65 0.625 0.6 0.575 0.55 0.525 0.5 M 2 admittance amplitude+trend 1920 1940 1960 1980 2000 M 2 amplitude: trend +18.6 yr cycle 1920 1940 1960 1980 2000 0.14.135 0.13.125 S 2 amplitude: trend + 18.6 yr cycle 270 M 2 phase 260 250 0.12 1900 1920 1940 1960 1980 2000 240 1920 1940 1960 1980 2000
Effects of the Noisy SF Record SF tides are quite noisy The analysis works well, despite methodological difficulties, because: The changes are large in M 2, where noise is worst There is a very long record Method 1 (right) takes out the 18.6 yr cycle using the admittance Method 2 (left) fits a trend plus an 18.6 yr cycle Both methods work pretty well for K 1, where noise is not too bad
Effects of the Noisy Record (Continued) M 2 is more difficult, but change is very large Note large residuals in strong ENSO periods, like 1939-42, 1982-83, 83, 1996-1999 1999
Effects of record Length (LOR) at SF This shows what happens when all 28, 37, 47 and 56 yr subsets of data are used Trends (mm/100yrs) and 95% conf limits as f[central time] Trend is estimated for each subset K 1 M 2 Each point is cen- tered at mid-pt of subset of data 28 yrs 28 yrs 37 yrs 37 yrs 47 yrs 47 yrs 56 yrs 56 yrs
Why the Observed Changes in SF Bay Tides? Large-scale changes Local causes of increased tides: Reduced flow, diking and channelization throughout bay ~50% reduction in river flow M 2 is increasing much faster than K 1 : may indicate that shelf stratification and internal tides have changed Noise in strong ENSO years may also show effects of local shelf water properties Local MSL rise may drive a change in resonance or reflection. We see effects of altered river flow right down to the entrance, which implies reflection.
Long-Term Changes in Seattle are Small 1920 1940 1960 1980 2000 Year Record extends from 1899-2008 MSL rising at ~0.2m/100 yrs, but tides are ~constant Small K 1 and M 2 trends are hard to separate from variability at 8 to 19 yrs 18.6yr nodal cycle does NOT follow astronomical forcing Both methods (fitting admittance and fitting amplitudes) have problems, especially for K 1 K 1 amplitude K 1 admittance ~9yr signal Residual ~19yr signal Residual M 2 amplitude M 2 residual 2.2 2.1 MSL Trend Mean Sea Level Residual Residual m 2 1.9
Methods Test: The influence of LOR on Trends at Seattle Seattle has LOR=109yr; use to test methods Look at all contiguous 28, 37, 47 and 56 yr subsets of admittance: Inferred trend varies with time and LOR used Weak trends that are often not significant Better removal of 18.6 yr variance might yield significant trends for K 1 Morals: Need long records! Need better methods! If no trend can be deduced from a 100+ yr record, then changes are small Trends (mm/100yrs) and 95% conf limits as f[central time] K 1 28 yrs 37 yrs 47 yrs 56 yrs M 2
Why aren t t Tides Changing at Seattle?? Seattle is a long ways from the ocean Are local density effects dominant? If local density effects are dominant, the deep waters of Puget Sound and St of Juan De Fuca may not be changing much, or Perhaps changes in the inland waters are compensating those in the coastal ocean None of the gauges around Puget Sound show much change But gauges in B.C. (e.g., Vancouver, Patricia Bay and Victoria) show moderate to rapid growth
Long-Term Changes at Astoria: K 1 Record extends from 1925-2008 2008 Local MSL rate ~-0.02m/100yrs ~ (not statistically significant) local tectonics and MSL rise are matched K 1 tide is also larger and earlier Amplitude, 35 ±4 4 mm/100 yrs Phase, -5 ±0.7º/100 yrs Relatively clean signal Largest K 1 change of any station 0.32 0.31 0.3 0.29 0.28 0.27 0.26 0.25 0.475 0.45 0.425 0.4 0.375 0.35 0.325 Trend in K 1 admittance amplitude 1930 1940 1950 1960 1970 1980 1990 2000 p K 1 amplitude: trend + 18.6 yr cycle 1930 1940 1950 1960 1970 1980 1990 2000 m 1.5 1.45 1.4 1.35 1.3 1.25 MSL Trend Mean Sea Level 1940 1960 1980 2000 Year deg 105 100 95 90 85 80 75 K 1 phase 1930 1940 1950 1960 1970 1980 1990 2000
Long-Term Changes at Astoria: M 2 M 2 tide is larger and earlier: Amplitude, 74 ±7 7 mm/100 yrs large! Phase, -8 ±0.5º/100 yrs (15 min) Tidal range: increasing ~1 ft/100yrs! S 2 is changing more rapidly than M 2 : implies reduced friction M 4 is also decreasing; implies reduced friction 0.24 S 2 amplitude: trend + 0.235 18.6 yr cycle 0.23 0.225 0.22 0.215 0.21 1930 1940 1950 1960 1970 1980 1990 2000 0.925 0.9 0.875 0.85 0.825 0.8 0.775 deg 1 0.95 0.9 0.85 Trend +M 2 admittance amplitude 1930 1940 1950 1960 1970 1980 1990 2000 Y 1.05 M 2 amplitude: trend + 18.6 yr cycle 47.5 45 42.5 40 37.5 35 32.5 30 1930 1940 1950 1960 1970 1980 1990 2000 Y M 2 phase 1930 1940 1950 1960 1970 1980 1990 2000
Why the Observed Changes in Amplitudes and Phases at Astoria? Astoria is notable for large changes in K 1, M 2 and S 2 Several factors: Regional increases throughout the NE Pacific Reduced friction due to: A narrower, deeper channel Loss of intertidal areas Reduced river flow has reduced friction Bed has degraded due to loss of sediment input likely has also reduced friction Likely not related to internal tides, because both M 2 and K 1 are growing rapidly Charleston and Newport are changing 40-50% as fast as Astoria: Probably represent the regional change w/o much human component 50-60% of the change at Astoria is due to internal changes to system
Conclusions Surprisingly large changes are occurring in the NE Pacific regional tidal regime Causes are complicated Local changes to individual harbors affect records There is a large-scale component due to changes in open-ocean ocean and/or shelf dynamics Open ocean part is likely related to climate change Possibly a useful indicator of climate change? Numerous ecosystem impacts; affects Coastal and estuarine habitats Navigation Coastal innundation The National Science Foundation has funded a continuation Analyze data for the entire Pacific What is the pattern of changes? Is there any acceleration of changes in tides? Numerical modeling to understand causes