Corrected calculations for vertical interpolation in Env-DATA

We have made a correction to the way we calculate vertical interpolations in Env-DATA. Vertical interpolation is available as an option for select variables in the weather models (ECMWF, NARR, and NCEP-DOE Reanalysis 2) when the tracking data contain information about the animal’s height in the attribute “height above ellipsoid” or "height above msl".

Background: The height of an animal or an estimate of wind speed can be given as the height above geoid (height above mean sea level) or the height above ellipsoid (a smooth surface approximating the geoid). We previously assumed that the geopotential values provided in the weather models gave a height above ellipsoid when divided by 9.80665. However, we recently discovered that this calculation actually provides the height above the geoid. In our corrected calculation, we now convert the animals’ “height above ellipsoid” values to the “height above msl”, and use these values for vertical interpolation.

The correction should make only a small difference in resulting values. The difference between geoid and ellipsoid height are typically less than 50 m, which is less than the vertical resolution of the models. (A pressure difference of 25 hpa, which is the smallest step size in any product, corresponds to roughly 200 meters.) Note that GPS units may provide elevations as the height above ellipsoid or the height above msl, and this is often not stated in the documentation. For example, GPS data from CTT include height above geoid/msl, while GPS data from e-obs include height above ellipsoid. If you are unsure what reference height is used for elevation values in your data, please contact your tag provider. Make sure that your values are stored in the correct variable in Movebank (height-above-ellipsoid or height above msl) so that vertical interpolations are calculated correctly.


Published by scd on 2015-02-06 07:41