AWIS Quality Control Procedures
Our quality control is done 365 day a year, on all weekends and all
holidays. It consists of three steps:
-
Automated quality control procedures screen the data before they enter
the database. These procedures detect suspicious dates, missing weather
parameters and values outside reasonable ranges. They ensure, for
example, that the maximum temperature is higher than the minimum
temperature and perform other statistical tests. Nearly all near real
time weather observations are sent in a complex code. Hourly data is
transmitted in METAR code while daily observations are sent in SHEF
code. Errors are frequently made in these coding systems. AWIS has
automated the decoding and checking of METAR and SHEF observations and
deals with most of the commonly made mistakes.
-
In the next step, we use color-coded map displays to visually inspect for
values that are outside of reasonable ranges (these are called
outliers). Can you spot the outlier in
this graphic of Kansas? Yes, this a manual step that takes time
and experience to ensure that you get the best weather data
from AWIS.
-
The cleansed data are then used to create gridded fields of each weather
parameter. Estimates from the gridded output are compared to the actual
observations. Discrepancies are manually examined and corrected data are
put into our database. Again, this is a manual step that helps us
provide our clients with the best data available.
Thousands of products are created after the above procedures
are completed. If clients require products before quality control
procedures are completed, preliminary products can be created and
delivered. Updates are delivered after the quality control is
completed.
Missing hourly data are estimated automatically when a single hour is
missing. If several hours are missing the data may be estimated by a
meteorologist. Missing parameters may also be estimated from gridded
output.
|
