Saturday, February 12, 2011

Data Processing and Tabulation of the Tide

NOAA provides a critical service to those who rely on the sea for their livelihood, such as commercial fishermen and those engaged in marine commerce.  Tides and currents are also critical for coastal zoning engineers, habitat restorers, ecologists, oceanagraphers, and atmospheric scientists.  Let's not forget those just out to enjoy the ocean or the bays in swimming, surfing, boating, etc.  NOAA takes its job very seriously and employs stringent quality control of its data, from the quality of the equipment used to the process of evaluating and reporting the results.

Tides and currents can also play a critical evidentiary role in a capital murder case such as this one.  Therefore, it's important that we understand how to use NOAA's tidal data and which station to use.

Which Station to Use
First, which station to use.  In the report I cited a few days ago, "Western Stege Marsh Restoration Project Year 5 Hydrologic Monitoring Report,"  the Richmond Station ID 9414863 is noted as the one to use for the Western Stege Marsh area, which includes the Conner Recovery Site:
Throughout the Year 5 monitoring period, continuous tidal elevations were measured by the National Oceanic and Atmospheric Administration (NOAA) at their Richmond tide gauge (NOAA gauge number 9414863).  Water levels measured at the NOAA gauge are representative of the adjacent bay and Meeker Slough. (pgs 1-2)
This map provided by the NOAA shows all of its stations in the Bay.


This map from Google shows the relationship of the Conner Recovery Site (blue tack) and the Laci Recovery Site (yellow tack) to both the NOAA Richmond station (green tack) and the San Francisco station (purple tack).  The Richmond station is obviously much closer to the two recovery sites than the San Francisco station.


The need to use the Richmond station data instead of the San Francisco station when analyzing conditions on the CRS and the LRS is proven by the NOAA's prediction tables.  While the NOAA has a monitoring station located at Richmond, Richmond's tidal predictions are based on the San Francisco station (the reference station), as are the tidal predictions for the Richmond Inner Harbor (CRS) and Point Isabel (LRS).  These are the corrections given at the top page of the 2003 predictions.  Note:  The Richmond monitoring station is located on the Chevron Oil Company Pier.
2003 NOAA Tide Predictions: Chevron Oil Company Pier, Richmond
(Reference station: San Francisco, Corrections Applied: Times: High +0 hr. 24 min., Low +0 hr. 38 min., Heights: High *1.04, Low *0.98)
2003 NOAA Tide Predictions: Richmond Inner Harbor
(Reference station: San Francisco, Corrections Applied: Times: High +0 hr. 24 min., Low +0 hr. 36 min., Heights: High *1.03, Low *0.98)
2003 NOAA Tide Predictions: Point Isabel
(Reference station: San Francisco, Corrections Applied: Times: High +0 hr. 23 min., Low +0 hr. 33 min., Heights: High +0.1, Low +0.0)

At the Richmond Station, as well as the Richmond Inner Harbor, the high tide peaks 24 minutes later than at the San Francisco station, and at Point Isabel, it's 23 minutes later.  An even greater time difference is noted for the peak low tide.  As for water level, Richmond's high tide will peak at 104% of the San Francisco station, and Richmond Inner Harbor at 103%.  Attempting to use the San Francisco station for times of tides and water levels would result in large errors.  It appears from the information he gave at trial that Dr. Cheng relied on the San Francisco station for his calculations.  


Which water level data to use

Predictions
The NOAA publishes prediction tables for each location based on astronomical factors, which don't include current meteorological conditions, such as rainfall or drought.  Predictions are for long-term planning.  And long-term planning is what is needed if anyone wants to take a field trip to observe the Recovery sites at water levels comparable to April 13 and April 14, 2003 -- because the tides are on a 25-hour cycle, the high tides and low tides occur at different times each day, and because the lunar cycle produces quite different water levels for the neap tides (1st and 3rd quarters of the moon cycle) than for the spring tides (full moons and new moons).  Obviously Rick Distaso, Dave Harris, and Craig Grogan didn't do their homework, as People's 100 was taken at a water level of 6.39 feet (link) and presented to the Court and Jury as representative of the Conner Site on April 13, 2003, when the water level that day reached only 5.88 ft.

Preliminary Water Levels
On its homepage, the Richmond station provides current information, which is about 20 minutes behind real-time.


Richmond also provides preliminary water levels in plot and table format, compared to the predictions.  This is a sampling of the table format. The Acoustc measurement has proven to be quite reliable for past field trips.

Tide Data

Station Date      Time  Pred 6 Acoustc  Backup
DCP#:                        1       1       2
Units:                    Feet    Feet    Feet
Data%:  MLLW     Local  100.00  39.375   0.000  
Maximum:                  5.68    5.34    0.00  
Minimum:                  1.01    1.11    0.00  
------- -------- ----- ------- ------- ------- 
9414863 20110210 00:00    3.35    3.19        
9414863 20110210 00:06    3.42    3.26        
9414863 20110210 00:12    3.49    3.32        


This is a current plot, showing the preliminary water level compared to the predicted.


When selecting your plot or table, be sure to choose local time.  

Verified Water Levels
Because of the time it takes for quality control, verified water levels aren't available for the Richmond station for about 10 days.  The following compares the preliminary water levels for Dec 22, 2010 with the verified for the peak high tide.  You can see that the Acoustc levels are exactly the same as the Verified.

Station Date Time Pred 6 Acoustc Backup
Maximum: 6.95 7.61
Minimum: -1.19 -0.75  
9414863 20101222 11:30    6.94    7.55        
9414863 20101222 11:36    6.95    7.57        
9414863 20101222 11:42    6.95    7.58        
9414863 20101222 11:48    6.94    7.61        
9414863 20101222 11:54    6.93    7.60        
9414863 20101222 12:00    6.91    7.57 [link]
__________________________________________
Station Date      Time  Pred 6  Vrfy 6
Maximum:                  6.95    7.61  
Minimum:                 -1.19   -0.75   
9414863 20101222 11:30    6.94    7.55
9414863 20101222 11:36    6.95    7.57
9414863 20101222 11:42    6.95    7.58
9414863 20101222 11:48    6.94    7.61
9414863 20101222 11:54    6.93    7.60
9414863 20101222 12:00    6.91    7.57 [link]


For December 23, the other day of my last field trip, the Verified 6-minute data has also confirmed the accuracy of the Acoustc data:


Station Date      Time  Pred 6 Acoustc  Backup
Maximum:                  6.77    6.90  
Minimum:                 -1.05   -0.89   
9414863 20101223 12:12    6.75    6.88        
9414863 20101223 12:18    6.76    6.90        
9414863 20101223 12:24    6.77    6.89        
9414863 20101223 12:30    6.77    6.88        
9414863 20101223 12:36    6.76    6.89        
9414863 20101223 12:42    6.75    6.90        
9414863 20101223 12:48    6.73    6.87 [link]

___________________________________________
Station Date Time Pred 6 Vrfy 6
Maximum: 6.77 6.90  
Minimum: -1.05 -0.89

9414863 20101223 12:12    6.75    6.88
9414863 20101223 12:18    6.76    6.90
9414863 20101223 12:24    6.77    6.89
9414863 20101223 12:30    6.77    6.88
9414863 20101223 12:36    6.76    6.89
9414863 20101223 12:42    6.75    6.90
9414863 20101223 12:48    6.73    6.87 [link]


The NOAA also converts the 6-minute verified data into other useful formats:  Hourly WL, High/low WL, and Monthly WL.  This is the High/low WL for December 22, 2010:


Station Date      Time Vrfy HL
Maximum:                  7.59  
Minimum:                 -0.74  
------- -------- ----- ------- 
9414863 20101222 01:18    5.89  H
9414863 20101222 05:42    3.73  L
9414863 20101222 11:48    7.59  HH 
9414863 20101222 18:54 -0.74 LL

You will notice that the HH for the day, which I underlined, is not the same as in the verified 6-minute data:  7.59 compared to 7.61.  Although it's only a difference of .02 feet, or slightly less than 1/4 of an inch, to be precise, the 6-minute water level should be used.  Sometimes there is more of a difference than that, and the time may even be different by a few minutes.  That is because “the times and heights of the high and low waters [are] derived from appropriate curve-fitting of the 6-minute interval data” ("Water Level Station Specifications and Deliverables For Shoreline Mapping Projects," May 2009   p. 17).

Storm Surge
Some have suggested that the storm surge from the storm on April 12, 2003, raised the water level at the Conner Recovery Site sufficiently to allow him to wash over the rocks.  Storm surge is reflected in the tide data from the Richmond Station.  Here is the plot for April 11-13, 2003.  The storm surge is the difference between the red line (verified water level) and the blue line (predicted water level).

The verified 6-minute water level table shows that the tide on April 13, 2003, was predicted to be only 5.45 but instead reached 5.88, a rise of 5.16 inches.

Station Date      Time  Pred 6  Vrfy 6
DCP#:                        1       1
Units:                    Feet    Feet
Data%:  MLLW     Local  100.00  100.00  
Maximum:                  5.45    5.88  
Minimum:                 -0.17    0.11 
9414863 20030413 09:36    5.44    5.86
9414863 20030413 09:42    5.45    5.87
9414863 20030413 09:48    5.45    5.83
9414863 20030413 09:54    5.45    5.84
9414863 20030413 10:00    5.44    5.86
9414863 20030413 10:06    5.43    5.88
9414863 20030413 10:12    5.41    5.88 [link]

Here is the High/low data for April 13, 2003.  You can see that it records the HH as 5.87 and the time as 10:00, whereas the 6-minute data shows that the water level peaked at 5.88 at 10:06.  As noted above, to be most accurate, the 6-minute data should be used.


Station Date      Time Vrfy HL
DCP#:                        1
Units:                    Feet
Data%:  MLLW     Local   1.667  
Maximum:                  5.87  
Minimum:                  0.12  
------- -------- ----- ------- 
9414863 20030413 04:00    2.39  L
9414863 20030413 10:00    5.87  HH
9414863 20030413 16:24    0.12  LL
9414863 20030413 23:12    5.68  H  [link]



Conclusion

For any analysis of tidal conditions on any given day at the Conner Recovery Site and the Laci Recovery Site,  a researcher must use the tide data from the Richmond Station ID 9414863.

Prediction tables are very useful for long-term planning, but cannot be relied upon in any analysis of tidal conditions on any given day.

Preliminary water level data (Acoustc) is available in about a 20 minute lag from real time.

Verified 6-minute water level data is available about 10 days later.  Even though Preliminary water levels are proving to be quite accurate, verified water levels should be obtained.

6-minute water level data, whether Preliminary (Acoustc) or Verified, must be used instead of the High/low data, as the latter may not precisely reflect the exact water level or the exact time it was reached.

No comments: