Skip to content

LMA errors and detection efficiency model based on input station locations and thresholds

License

Notifications You must be signed in to change notification settings

vbalderdash/LMAsimulation

Repository files navigation

Original release of model in conjunction with publication:

DOI

This model and its results are discussed in an article avaiable through the Journal of Geophysical Research (http://onlinelibrary.wiley.com/doi/10.1002/2016JD025159/full).

Please cite:

V. C. Chmielewski and E. C. Bruning (2016), Lightning Mapping Array flash detection performance with variable receiver thresholds, J. Geophys. Res. Atmos., 121, 8600-8614, doi:10.1002/2016JD025159

If any results from this model are presented.

LMAsimulation

This code is designed to simulate the way any LMA would solve a grid of point emitters, accounting for each station's noise levels and adding normally distributed timing errors at each station.

LMAsimulation_full.ipynb

This is the primary notebook for plotting the errors in the solutions over a regular, gridded domain in a Monte Carlo model. The used stations and thresholds can be taken from a time range of station log files (with any set of stations exempted) or from a list of stations in the network.csv file.

The grid of points is set in 'initial_points' in x,y,z in the map projection plane. The solutions are currently all found through scipy's lstsq function. The function can also be run for several iterations at a singular location.

The flash detection efficiency is calculated from the WTLMA climatology results in fde.csv. The flash area error estimate is calculated from the typical_flashes.csv file which contains the median flash size and number of points of flashes with at least x+9 number of points (per line).

Note: The log file reading function is set up to handle errors in files through Linux subprocesses, so any errors may cause it to crash on other platforms. The errored logs are copied to an '*_original' file and the bad lines are removed from the active file.

CurvatureMatrix.ipynb

This runs the curvature matrix calculations for rmse of solution points in a gridded domain. The used stations must be specified in the network.csv file.

LMAdetection_efficiency.ipynb

Stand alone notebook purely for quickly estimating the detection efficiency of a network.

This is a new addition not described in the publication.

LMAsimulation_ellipse.ipynb

Based on the same theory as the full simulation, but instead returns the expected errors by x-y covariances and z standard deviations. Otherwise the simulation is the same, but resulting plots from the covariances are more intuitive for larger/combined networks where the center of the contributing set of stations may be displaced from the network center resulting in non-radial error patterns.

simulation_function.py

This contains the calculations for all models.

simulation_ellipse.py

Based on simulation_function.py but with extra functionality for calculating the error covariances in x and y.

parsed_functions.py

This contains only the calculations needed for the simplified detection efficiency model used by LMAdetection_efficiency.ipynb and LMAsimulation_ellipse.ipynb

read_logs.py

This is intended for parsing, quick QC, and integrating v10 log files for LMA stations.

coordinateSystems.py

This contains the driving functions for the coordinate system transformations

network_full.csv

This csv contains the approximated station locations and thresholds used for analysis. For more precise station locations, please contact the inidividual network operators or the authors of this program.

About

LMA errors and detection efficiency model based on input station locations and thresholds

Resources

License

Stars

Watchers

Forks

Packages

No packages published