Transcript Document
DATA QUALITY CONTROL
Julie Thomas
Coastal Data Information Program (CDIP)
Scripps Institution of Oceanography
CI Design Workshop
17-19 October 2007
La Jolla, CA
ACT
QARTOD
REAL-TIME CASE STUDY
INTERNATIONAL EFFORTS
WMO, OceanSITES/Eurosites/IFREMER, JCOMM, ……..
ALLIANCE FOR COASTAL TECHNOLOGIES
ACT
Wave Sensor Technologies
University of South Florida, March 7-9 2007
By Bill O’Reilly
Directional Wave Observations
1) The Basics
2) Remote Sensing v. In Situ : It’s all good.
3) Estimating directional wave properties.
4) Measurement accuracy and user needs.
5) A Recipe for instrument evaluation.
The Basics: Estimating the Motion of a Sea Surface Particle
z
z
dz/dx,
dz/dy
y
x
w
The Big 3
v
u
X, Y, Z
Pressure Sensors
Accelerometers
Tilt sensors
Angular Rate Sensors
Acoustic Sensors
Radar
Lidar
In-Situ and Remotely Sensed Waves
Method
O(cm) X,Y,Z
Accuracy
Spatial
Coverage
Time
Coverage
In situ
SAR,LIDAR
HF, X-Band Radar
In situ and remote sensing systems are complimentary, NOT redundant.
XYZ TIME SERIES
The Big 3: X, Y, Z Time Series Analysis The First 5: S(f),a1(f),b1(f),a2(f),b2(f) !!
S
Theoretical maximum level of breaking wave in a fully developed sea
Each box represents the number of hourly observation – Datawell
Buoy Mark II w/ hippy used as data source (standard).
GPS instrument dropped out during low
Frequency – causing spikes in data
Instruments tend to disagree as
they are interpreting sea surface
differently under breaking wave
conditions.
Comparison between a Datawell Buoy w/ hippy and a PROTOTYPE
Datawell w/ GPS. Low numbers indicate good correlations. Instruments
Were co-located.
Datawell w/ hippy and Datawell w/ GPS tend to agree well with wave direction.
Note how both instruments reach their noise floor during long period, low energy
events.
GPS antenna on top of the buoy has a “natural”
noise when it wobbles – a method has been
devised to correct for this problem.
Directional spread is where many instruments fall apart! Good indication of the
quality of the instrument – demonstrates their directional noise level!
QUALITY ASSURANCE OF REAL-TIME DATA
QARTOD
http://qartod.org
QARTOD is a continuing multi-agency effort to address the Quality Assurance
and Quality Control issues of the Integrated Ocean Observing System (IOOS)
and broader international community.
Quality Control standards have been submitted to DMAC for both waves and
Currents. Ocean.US is now soliciting submissions from the Regional
Associations for all standards including QA/QC. The QARTOD submissions
Will be posted soon on the Ocean.US website.
Parameters addressed through QARTOD: Waves, In-situ Currents, Remote
Currents (HF RADAR), CTDs, Dissolved Oxygen.
Collaborated with DMAC Metadata team for a one day appended workshop
During QARTOD IV, June 2006.
Quality Assurance (QA) –
Verify that instrumentation is
calibrated and tested to assure
collection of the highest quality
data possible.
Quality Control (QC) –
Analyze and verify the data stream
to assure the highest quality data
possible.
http://qartod.org
http://cdip.ucsd.edu/documents/index/product_docs/qc_summaries/waves/waves_table.php
VALIDATION CASE STUDY - Real-time Buoy Data
OREGON
Humboldt Bay South Spit (128)
CALIFORNIA
Eel River
Cape Mendocino (094)
NDBC BUOY 46022, EEL RIVER, re-deployed this summer
with 3DMG sensor. 46022 is 34 miles away from buoy
46213 (CDIP Datawell Buoy Station 094).
Full spectra looks ok at first glance!
Noisy directional spread
Under predicts LONG PERIOD SWELL (12-33 sec)
46022 under predicts low frequency waves.
Buoys are moored 34 miles apart.
Part of this difference in the high
frequency could be caused by local
sea conditions.
Comparison of the Datawell /hippy and the NDBC 46022 (Eel River Buoy)
w/ 3DMG sensor .
Why is directional spread important? Why do
we need accurate wave measurements?
The models are only
as good as their input
source.
In Southern California,
for instance, 5
degrees can make
a big difference in how
much energy reaches
the beach.
Let’s come together as an IOOS community and further validate
our wave instrumentation!