EVA - Extrem Values Analysis
============================


Annual maxima
-------------

.. code::

    from warsa.discharge import eva

The function ``get_annual_maxima`` has as a series (not a data frame) input parameter with datetime as index and discharges as values.

.. code::

    sr_annual_maxima = eva.get_annual_maxima(sr)

The output is a pandas series with year as index and the maxima as value::

    1979     4.078448
    1980    10.036425
    1981    15.106880
    ...
    2012     7.410849
    2013    22.590010
    2014    10.074438

Simple plot:

.. code::

    sr_annual_maxima.plot(kind='bar', color='b', title='Maximal annual discharge starting on 1st Nov.')
    plt.show()

.. image:: images/q_annual_maxima_Jan01.png


If not otherwise stated, the year starts on 1st January at 00:00:00. Hydrological years normally start in another month and at another time, e.g., on 1st November at 07:30am in Germany. In order to get maxima from 1st November 07:30 to 31th October before 07:30 of the next year:

.. code::

    sr_annual_maxima = eva.get_annual_maxima(sr, datetime.datetime(2000, 11, 1, 7, 30))

In the example above, the year is only a place holder. The function will split all years from YYYY-11-01 07:30:00 to (YYYY+1)-10-31 07:29:59.99999

The output is a pandas series with year as index and the maxima as value::

    1980     6.514527
    1981    15.106880
    1982    11.751186
    ...
    2012     5.625469
    2013    22.590010
    2014    10.074438
    2015     5.337633

Note the difference between both outputs.


Plot:

.. code::

    ax = sr_annual_maxima.plot(kind='bar', color='b', title='Maximal annual discharge starting on 1st Nov.')
    ax.set_xlabel("Year")
    ax.set_ylabel("Qmax (m³/s)")
    plt.tight_layout()
    plt.show()


.. image:: images/q_annual_maxima_Nov01.png