DAPPER is a set of templates for benchmarking the performance of data assimilation (DA) methods. The numerical experiments provide support and guidance for new developments in DA. The typical set-up is a synthetic (twin) experiment, where you specify a dynamic model and an observational model, and use these to generate a synthetic truth (multivariate time series), and then estimate that truth given the models and noisy observations.
- Read & run examples/
basic_1.py
andbasic_2.py
, or their corresponding notebooks (requires Google login). - This screencast provides an overview to DAPPER.
- Install.
- The documentation includes general guidelines and the API reference, but most users must expect to read the code as well.
- If used towards a publication, please cite as The experiments used (inspiration from) DAPPER [ref], version 1.6.0, or similar, where [ref] points to .
- Also see the interactive tutorials on DA theory with Python.
DAPPER enables the numerical investigation of DA methods
through a variety of typical test cases and statistics. It
(a) reproduces numerical benchmarks results reported in the literature, and
(b) facilitates comparative studies, thus promoting the
(a) reliability and
(b) relevance of the results.
For example, the figure below is generated by docs/examples/basic_3.py
,
reproduces figure 5.7 of these lecture notes.
DAPPER is
(c) open source, written in Python, and
(d) focuses on readability;
this promotes the
(c) reproduction and
(d) dissemination of the underlying science,
and makes it easy to adapt and extend.
DAPPER demonstrates how to parallelise ensemble forecasts (e.g., the QG model),
local analyses (e.g., the LETKF), and independent experiments (e.g., docs/examples/basic_3.py
).
It includes a battery of diagnostics and statistics,
which all get averaged over subdomains (e.g., "ocean" and "land") and then in time.
Confidence intervals are computed, including correction for auto-correlations,
and used for uncertainty quantification, and significant digits printing.
Several diagnostics are included in the on-line "liveplotting" illustrated below,
which may be paused for further interactive inspection.
In summary, DAPPER is well suited for teaching and fundamental DA research.
Also see its drawbacks.
Successfully tested on Linux/Mac/Windows.
If you're an expert, setup a python environment however you like. Otherwise: Install Anaconda, then open the Anaconda terminal and run the following commands:
conda create --yes --name dapper-env python=3.12
conda activate dapper-env
python --version
Ensure the printed version is as desired. Keep using the same terminal for the commands below.
Do you want the DAPPER code available to play around with? Then
- Download and unzip (or
git clone
) DAPPER. - Move the resulting folder wherever you like,
andcd
into it (ensure you're in the folder with asetup.py
file). pip install -e '.'
Do you just want to run a script that requires DAPPER? Then
- If the script comes with a
requirements.txt
file that lists DAPPER, then do
pip install -r path/to/requirements.txt
. - If not, hopefully you know the version of DAPPER needed. Run
pip install dapper==1.6.0
to get version1.6.0
(as an example).
You should now be able to do run your script with
python path/to/script.py
.
For example, if you are in the DAPPER dir,
python docs/examples/basic_1.py
PS: If you closed the terminal (or shut down your computer),
you'll first need to run conda activate dapper-env
Method | Literature reproduced |
---|---|
EnKF 1 | Sakov08, Hoteit15, Grudzien2020 |
EnKF-N | Bocquet12, Bocquet15 |
EnKS, EnRTS | Raanes2016 |
iEnKS / iEnKF / EnRML / ES-MDA 2 | Sakov12, Bocquet12, Bocquet14 |
LETKF, local & serial EAKF | Bocquet11 |
Sqrt. model noise methods | Raanes2014 |
Particle filter (bootstrap) 3 | Bocquet10 |
Optimal/implicit Particle filter 3 | Bocquet10 |
NETF | Tödter15, Wiljes16 |
Rank histogram filter (RHF) | Anderson10 |
4D-Var | |
3D-Var | |
Extended KF | |
Optimal interpolation | |
Climatology |
1: Stochastic, DEnKF (i.e. half-update), ETKF (i.e. sym. sqrt.).
Serial forms are also available.
Tuned with inflation and "random, orthogonal rotations".
2: Also supports the bundle version,
and "EnKF-N"-type inflation.
3: Resampling: multinomial
(including systematic/universal and residual).
The particle filter is tuned with "effective-N monitoring",
"regularization/jittering" strength, and more.
For a list of ready-made experiments with suitable,
tuned settings for a given method (e.g., the iEnKS
), use:
grep -r "xp.*iEnKS" dapper/mods
Simple models facilitate the reliability, reproducibility, and interpretability of experiment results.
Model | Lin | TLM** | PDE? | Phys.dim. | State len | Lyap≥0 | Implementer |
---|---|---|---|---|---|---|---|
Id | Yes | Yes | No | N/A | * | 0 | Raanes |
Linear Advect. (LA) | Yes | Yes | Yes | 1d | 1000 * | 51 | Evensen/Raanes |
DoublePendulum | No | Yes | No | 0d | 4 | 2 | Matplotlib/Raanes |
Ikeda | No | Yes | No | 0d | 2 | 1 | Raanes |
LotkaVolterra | No | Yes | No | 0d | 5 * | 1 | Wikipedia/Raanes |
Lorenz63 | No | Yes | "Yes" | 0d | 3 | 2 | Sakov |
Lorenz84 | No | Yes | No | 0d | 3 | 2 | Raanes |
Lorenz96 | No | Yes | No | 1d | 40 * | 13 | Raanes |
Lorenz96s | No | Yes | No | 1d | 10 * | 4 | Grudzien |
LorenzUV | No | Yes | No | 2x 1d | 256 + 8 * | ≈60 | Raanes |
LorenzIII | No | No | No | 1d | 960 * | ≈164 | Raanes |
Vissio-Lucarini 20 | No | Yes | No | 1d | 36 * | 10 | Yumeng |
Kuramoto-Sivashinsky | No | Yes | Yes | 1d | 128 * | 11 | Kassam/Raanes |
Quasi-Geost (QG) | No | No | Yes | 2d | 129²≈17k | ≈140 | Sakov |
*
: Flexible; set as necessary**
: Tangent Linear Model included?
The models are found as subdirectories within dapper/mods
.
A model should be defined in a file named __init__.py
,
and illustrated by a file named demo.py
.
Most other files within a model subdirectory
are usually named authorYEAR.py
and define a HMM
object,
which holds the settings of a specific twin experiment,
using that model,
as detailed in the corresponding author/year's paper.
A list of these files can be obtained using
find dapper/mods -iname '[a-z]*[0-9]*.py'
Some files contain settings used by several papers.
Moreover, at the bottom of each such file should be (in comments)
a list of suitable, tuned settings for various DA methods,
along with their expected, average rmse.a
score for that experiment.
As mentioned above, DAPPER reproduces literature results.
You will also find results that were not reproduced by DAPPER.
DAPPER is aimed at research and teaching (see discussion up top). Example of limitations:
- It is not suited for very big models (>60k unknowns).
- Non-uniform time sequences.
The scope of DAPPER is restricted because
Moreover, even straying beyond basic configurability appears unrewarding when already building on a high-level language such as Python. Indeed, you may freely fork and modify the code of DAPPER, which should be seen as a set of templates, and not a framework.
Also, DAPPER comes with no guarantees/support. Therefore, if you have an operational or real-world application, such as WRF, you should look into one of the alternatives, sorted by approximate project size.
Name | Developers | Purpose (approximately) |
---|---|---|
DART | NCAR | General |
PDAF | AWI | General |
JEDI | JCSDA (NOAA, NASA, ++) | General |
OpenDA | TU Delft | General |
EMPIRE | Reading (Met) | General |
ERT | Statoil | History matching (Petroleum DA) |
PIPT | CIPR | History matching (Petroleum DA) |
MIKE | DHI | Oceanographic |
OAK | Liège | Oceanographic |
Siroco | OMP | Oceanographic |
Verdandi | INRIA | Biophysical DA |
PyOSSE | Edinburgh, Reading | Earth-observation DA |
Below is a list of projects with a purpose more similar to DAPPER's (research in DA, and not so much using DA):
Name | Developers | Notes |
---|---|---|
DAPPER | Raanes, Chen, Grudzien | Python |
SANGOMA | Conglomerate* | Fortran, Matlab |
hIPPYlib | Villa, Petra, Ghattas | Python, adjoint-based PDE methods |
FilterPy | R. Labbe | Python. Engineering oriented. |
DASoftware | Yue Li, Stanford | Matlab. Large inverse probs. |
Pomp | U of Michigan | R |
EnKF-Matlab | Sakov | Matlab |
EnKF-C | Sakov | C. Light-weight, off-line DA |
pyda | Hickman | Python |
PyDA | Shady-Ahmed | Python |
DasPy | Xujun Han | Python |
DataAssim.jl | Alexander-Barth | Julia |
DataAssimilationBenchmarks.jl | Grudzien | Julia, Python |
EnsembleKalmanProcesses.jl | Clim. Modl. Alliance | Julia, EKI (optim) |
Datum | Raanes | Matlab |
IEnKS code | Bocquet | Python |
The EnKF-Matlab
and IEnKS
codes have been inspirational
in the development of DAPPER.
*
: AWI/Liege/CNRS/NERSC/Reading/Delft
Do not hesitate to open an issue, whether to report a problem or ask a question. It may take some time for us to get back to you, since DAPPER is primarily a volunteer effort. Please start by perusing the documentation and searching the issue tracker for similar items.
Pull requests are very welcome. Examples: adding a new DA method, dynamical models, experimental configuration reproducing literature results, or improving the features and capabilities of DAPPER. Please keep in mind the intentional limitations and read the developers guidelines.
Patrick N. Raanes, Yumeng Chen, Colin Grudzien, Maxime Tondeur, Remy Dubois
DAPPER is developed and maintained at NORCE (Norwegian Research Institute) and the Nansen Environmental and Remote Sensing Center (NERSC), in collaboration with the University of Reading, the UK National Centre for Earth Observation (NCEO), and the Center for Western Weather and Water Extremes (CW3E).
- Combining data assimilation and machine learning to emulate a dynamical model from sparse and noisy observations: A case study with the Lorenz 96 model
- Adaptive covariance inflation in the ensemble Kalman filter by Gaussian scale mixtures
- Revising the stochastic iterative ensemble smoother
- p-Kernel Stein Variational Gradient Descent for Data Assimilation and History Matching
- Springer book chapter: Data Assimilation for Chaotic Dynamics