Hydropandas is a Python package for reading, analyzing and writing (hydrological) timeseries data.
The HydroPandas package allows users to store a timeseries and metadata in a single object. This object inherits from a pandas DataFrame, with all its wonderful features, and is extended with custom methods and attributes related to hydrological timeseries.
The HydroPandas package also provides convenient read functions for Dutch hydrological data from:
- BRO
- DINO
- FEWS PI-XML
- KNMI
- Lizard
- MODFLOW groundwater models
- IMOD groundwater models
- Pastastore
- Waterinfo
- WISKI csv files
Install the module with pip:
pip install hydropandas
HydroPandas requires pandas
, scipy
, matplotlib
, tqdm
, requests
and colorama
.
For some functionality additional packages are required:
geopandas
: for dealing with shapefilespastastore
: for reading or storing data from PastaStorebokeh
,branca
,folium
: for interactive mapsflopy
: for reading data from MODFLOW modelsxarray
: for loading data from REGIS
For installing in development mode, clone the repository and install by
typing pip install -e .
from the module root directory.
For installing all the optional packages use pip install -e .[full]
.
- Questions on HydroPandas ("How can I?") can be asked and answered on Github Discussions.
- Bugs, feature requests and other improvements can be posted as Github Issues.
- Find out how to contribute to HydroPandas at our Contribution page.
Importing a groundwater time series from the BRO using the BRO-id and the tube number:
import hydropandas as hpd
gw_bro = hpd.GroundwaterObs.from_bro("GMW000000041261", 1)
Or import all groundwater time series from the BRO within a certain extent:
oc = hpd.read_bro(extent=(117850, 118180, 439550, 439900))
The Obs class holds the measurements and metadata for one timeseries. There are currently 5 specific Obs classes for different types of measurements:
- GroundwaterObs: for groundwater measurements
- WaterQualityObs: for groundwater quality measurements
- WaterlvlObs: for surface water level measurements
- ModelObs: for "observations" from a MODFLOW model
- MeteoObs: for meteorological observations
- PrecipitationObs: for precipitation observations, subclass of MeteoObs
- EvaporationObs: for evaporation observations, subclass of MeteoObs
Each of these Obs classes is essentially a pandas DataFrame with additional methods and attributes related to the type of measurement that it holds. Each Obs object also contain specific methods to read data from specific sources.
The ObsCollection class, as the name implies, represents a collection of Obs classes, e.g. 10 timeseries of the groundwater level in a certain area. The ObsCollection is also a pandas DataFrame in which each timeseries is stored in a different row. Each row contains metadata (e.g. latitude and longitude of the observation point) and the Obs object (DataFrame) that holds the measurements. It is recommended to let an ObsCollection contain only one Obs type, e.g. to create an ObsCollection for 10 GroundwaterObs, and a separate ObsCollection for 5 PrecipitationObs.
Like the Obs class, the ObsCollection class contains a bunch of methods for reading data from different sources. See the next section for supported data sources.
- Onno Ebbens, Artesia
- Ruben Caljé, Artesia
- Davíd Brakenhoff, Artesia
- Martin Vonk, Artesia