 |
| Titel |
wradlib - an Open Source Library for Weather Radar Data Processing |
| VerfasserIn |
Thomas Pfaff, Maik Heistermann, Stephan Jacobi |
| Konferenz |
EGU General Assembly 2014
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 16 (2014) |
| Datensatznummer |
250091988
|
| Publikation (Nr.) |
 EGU/EGU2014-6308.pdf |
|
|
|
|
|
| Zusammenfassung |
| Even though weather radar holds great promise for the hydrological sciences, offering
precipitation estimates with unrivaled spatial and temporal resolution, there are still problems
impeding its widespread use, among which are:
almost every radar data set comes with a different data format with public reading
software being available only rarely.
standard products as issued by the meteorological services often do not serve the
needs of original research, having either too many or too few corrections applied.
Especially when new correction methods are to be developed, researchers are
often forced to start from scratch having to implement many corrections in
addition to those they are actually interested in.
many algorithms published in the literature cannot be recreated using the
corresponding article only. Public codes, providing insight into the actual
implementation and how an approach deals with possible exceptions are rare.
the radial scanning setup of weather radar measurements produces additional
challenges, when it comes to visualization or georeferencing of this type of data.
Based on these experiences, and in the hope to spare others at least some of these tedious
tasks, wradlib offers the results of the author’s own efforts and a growing number of
community-supplied methods.
wradlib is designed as a Python library of functions and classes to assist users in their
analysis of weather radar data. It provides solutions for all tasks along a typical processing
chain leading from raw reflectivity data to corrected, georeferenced and possibly gauge
adjusted quantitative precipitation estimates. There are modules for data input/output, data
transformation including Z/R transformation, clutter identification, attenuation correction,
dual polarization and differential phase processing, interpolation, georeferencing,
compositing, gauge adjustment, verification and visualization.
The interpreted nature of the Python programming language makes wradlib an ideal tool
for interactive data exploration and analysis. Based on the powerful scientific python stack
(numpy, scipy, matplotlib) and in parts augmented by functions compiled in C or Fortran,
most routines are fast enough to also allow data intensive re-analyses or even real-time
applications.
From the organizational point of view, wradlib is intended to be community driven. To
this end, the source code is made available using a distributed version control system (DVCS)
with a publicly hosted repository. Code may be contributed using the fork/pull-request
mechanism available to most modern DVCS. Mailing lists were set up to allow dedicated
exchange among users and developers in order to fix problems and discuss new
developments.
Extensive documentation is a key feature of the library, and is available online at
http://wradlib.bitbucket.org. It includes an individual function reference as well as examples,
tutorials and recipes, showing how those routines can be combined to create complete
processing workflows. This should allow new users to achieve results quickly, even without
much prior experience with weather radar data. |
|
|
|
|
|
|