Working with catalog-based forecasts

This example shows some basic interactions with data-based forecasts. We will load in a forecast stored in the CSEP data format, and compute the expected rates on a 0.1° x 0.1° grid covering the state of California. We will plot the expected rates in the spatial cells.

Overview:

1. Define forecast properties (time horizon, spatial region, etc).

2. Compute the expected rates in space and magnitude bins

3. Plot expected rates in the spatial cells

Load required libraries

Most of the core functionality can be imported from the top-level csep package. Utilities are available from the csep.utils subpackage.

import numpy

import csep
from csep.core import regions
from csep.utils import datasets

Load data forecast

PyCSEP contains some basic forecasts that can be used to test of the functionality of the package. This forecast has already been filtered to the California RELM region.

forecast = csep.load_catalog_forecast(datasets.ucerf3_ascii_format_landers_fname)

Define spatial and magnitude regions

Before we can compute the bin-wise rates we need to define a spatial region and a set of magnitude bin edges. The magnitude bin edges # are the lower bound (inclusive) except for the last bin, which is treated as extending to infinity. We can bind these # to the forecast object. This can also be done by passing them as keyword arguments into csep.load_catalog_forecast().

# Magnitude bins properties
min_mw = 4.95
max_mw = 8.95
dmw = 0.1

# Create space and magnitude regions
magnitudes = regions.magnitude_bins(min_mw, max_mw, dmw)
region = regions.california_relm_region()

# Bind region information to the forecast (this will be used for binning of the catalogs)
forecast.region = regions.create_space_magnitude_region(region, magnitudes)

Compute spatial event counts

The csep.core.forecasts.CatalogForecast provides a method to compute the expected number of events in spatial cells. This requires a region with magnitude information.

_ = forecast.get_expected_rates(verbose=True)

Processed 1 catalogs in 0.001 seconds
Processed 2 catalogs in 0.002 seconds
Processed 3 catalogs in 0.003 seconds
Processed 4 catalogs in 0.003 seconds
Processed 5 catalogs in 0.004 seconds
Processed 6 catalogs in 0.005 seconds
Processed 7 catalogs in 0.005 seconds
Processed 8 catalogs in 0.006 seconds
Processed 9 catalogs in 0.008 seconds
Processed 10 catalogs in 0.009 seconds
Processed 20 catalogs in 0.015 seconds
Processed 30 catalogs in 0.023 seconds
Processed 40 catalogs in 0.030 seconds
Processed 50 catalogs in 0.037 seconds
Processed 60 catalogs in 0.043 seconds
Processed 70 catalogs in 0.049 seconds
Processed 80 catalogs in 0.056 seconds
Processed 90 catalogs in 0.063 seconds
Processed 100 catalogs in 0.069 seconds
Processed 200 catalogs in 0.130 seconds
Processed 300 catalogs in 0.194 seconds
Processed 400 catalogs in 0.257 seconds
Processed 500 catalogs in 0.348 seconds
Processed 600 catalogs in 0.410 seconds
Processed 700 catalogs in 0.473 seconds
Processed 800 catalogs in 0.567 seconds
Processed 900 catalogs in 0.628 seconds
Processed 1000 catalogs in 0.691 seconds
Processed 2000 catalogs in 1.455 seconds
Processed 3000 catalogs in 2.170 seconds
Processed 4000 catalogs in 2.910 seconds
Processed 5000 catalogs in 3.633 seconds
Processed 6000 catalogs in 4.351 seconds
Processed 7000 catalogs in 5.092 seconds
Processed 8000 catalogs in 5.784 seconds
Processed 9000 catalogs in 6.557 seconds
Processed 10000 catalogs in 7.270 seconds

Plot expected event counts

We can plot the expected event counts the same way that we plot a csep.core.forecasts.GriddedForecast

ax = forecast.expected_rates.plot(plot_args={'clim': [-3.5, 0]}, show=True)

The images holes in the image are due to under-sampling from the forecast.

Quick sanity check

The forecasts were filtered to the spatial region so all events should be binned. We loop through each data in the forecast and count the number of events and compare that with the expected rates. The expected rate is an average in each space-magnitude bin, so we have to multiply this value by the number of catalogs in the forecast.

total_events = 0
for catalog in forecast:
    total_events += catalog.event_count
numpy.testing.assert_allclose(total_events, forecast.expected_rates.sum() * forecast.n_cat)