Tutorial notebooks

The tutorial lives in the notebooks/ directory of the repository. The six numbered notebooks replace the six DART_LAB slide sections:

  1. 01_da_concepts_in_1d – Bayes’ rule, the product of Gaussians, the Kalman filter, and ensemble filters (EAKF, EnKF, RHF) in one dimension.

  2. 02_multivariate_assimilation – regression of observation increments onto unobserved variables; Lorenz 63 and Lorenz 96.

  3. 03_inflation_and_localization – rank histograms, variance inflation, regression sampling error, and Gaspari-Cohn localization.

  4. 04_nongaussian_qceff – bounded variables, the rank histogram filter, quantile conservation, and probit (PPI) transforms.

  5. 05_adaptive_inflation – Bayesian inflation updates, spatially varying adaptive inflation, observing networks.

  6. 06_the_real_dart_system – running OSSEs with the Fortran DART system, namelist control, QCEFF tables and diagnostics.

00_getting_started checks your environment and introduces the interactive tools and the color conventions.

Apps

Each interactive tool matches a MATLAB DART_LAB app of the same name:

App

Teaches

gaussian_product

product of two Gaussians

oned_ensemble

EAKF / EnKF / RHF updates of a 1-D ensemble

oned_cycle

continuous Kalman filter vs ensemble filters

oned_model

cycling DA with model error

oned_model_inf

adaptive inflation in 1-D

twod_ensemble

updating an unobserved variable by regression

twod_ppi_ensemble

QCEFF / probit-transformed regression

bounded_oned_ensemble

filters for non-negative variables

run_lorenz_63

ensemble DA on the Lorenz 63 attractor

run_lorenz_96

localization and inflation in 40 variables

run_lorenz_96_inf

spatially varying adaptive inflation