Lab assignment 1

Due Friday, February 7

For lab assignment 1, we’ll be using various python tools that are useful for analyzing and visualizing weather data, including MetPy, developed by NSF Unidata. The first step, if you haven’t done this before, will be to install python, and then ultimately install MetPy and related packages.

Note on python installation

For some of the things we will do in class with python, you could quite easily use Google’s Colab to run the needed python code and notebooks. This could save you some time initially if you don’t have a lot of python experience or run into trouble. However, for activities later in the semester, it’ll likely be a lot easier to have it installed on your computer, because we’ll be dealing with larger files that aren’t saved in the cloud, etc. So I’d recommend following the install steps below, but will leave that choice up to you as to what you think will work best for your own preferences and computing situation.

Installing python via miniforge

First, let’s install the miniforge version of python. (If you already have miniconda/anaconda/mamba installed on the computer you want to use, you can skip this step and go to the assignment. Though you may still want to create a new conda environment for the class.) Following the instructions, which are based on those in the Unidata python workshop, but modified to use Miniforge/mamba instead, which is much faster:

Windows

• Download the Miniforge installer for Windows.

• Follow the prompts, taking note of the options to “Create start menu shortcuts” and “Add Miniforge3 to my PATH environment variable”. The latter is not selected by default due to potential conflicts with other software. Without Miniforge3 on the path, the most convenient way to use the installed software (such as commands conda and mamba) will be via the “Miniforge Prompt” installed to the start menu.

• You should now have a program called “Miniforge Prompt” installed. Open it (this will be your Python command prompt).

Mac/Linux

• Download the Miniforge bash installer for your platform.

• After downloading the bash installer, open a command prompt (terminal app on the Mac).

• Change the directory at the terminal to wherever the installer was downloaded. On most systems, this will default to the downloads directory in your user account. If that’s the case, cd ~/Downloads will get you there, or replace the path with wherever you saved the file.

• Run the installer script by typing bash Miniforge3-MacOSX-arm64.sh. Note: Your file name may be different depending upon your operating system! replace Miniforge3-MacOSX-arm64.sh with whatever the name of the file you downloaded was.

• Accept the defaults.

• After the installer has completed completely close and restart your terminal program (this sources the newly modified path).

• If bash isn’t your default shell, switch to it by running the command bash.

• Verify that your install is working by running mamba --version. You should see a response like mamba 1.5.6, conda 23.11.0 or similar (though yours may be a different version number).

Setting up your environment

Now we will set up an environment with the packages we need to have installed. Here is a link to an environment file that we’ll use for the class (adapted from Unidata’s workshop materials): environment.yml file

To set up this environment, follow these steps:

• Open a terminal window (Miniforge Prompt if you’re on Windows).

• Download the .yml file that tells your system what should be in the environment (see link above).

• In the terminal, navigate to wherever this file is saved, probably something like cd ~/Downloads will get you there.

• Run the command conda env create -f environment_ats641_2025.yml and wait for the installation to finish (it may take a while, especially if you’re on a slow internet connection).

• Run the command conda activate ats641_2025 to activate the environment you created and verify that everything is ready. (It may ask you to do something like conda init bash, if so then do that first.)

Opening and running a jupyter notebook

There are different ways you can run and interact with python, but a great way to get started is with Jupyter notebooks. They allow for you to write and test your code in a really user-friendly way.

• At the terminal, cd into whatever directory you want to work out of (this might be a directory you’ve set up just for the class, or you can make a new one, etc.)

• if you haven’t, run conda activate ats641_2025 to activate the environment.

• We’re going to start with an example notebook, obtained from the Unidata website. Right-click and download this file. (This notebook originates from this page.)

• Now, open Jupyter Lab, by simply running jupyter lab. This will open a new browser tab with Jupyter Lab in it. Or you can download the desktop app which I personally like to use.

• Click on ‘Python3’ to launch the python kernel.

• Open up the index.ipynb notebook (double-click it from the menu on the left) – this will give you a feel for what a notebook looks like and how it works.

• Walk through the steps in this notebook (Hint: you can run the code in cells using shift-R, or using the “play button” at the top) and hopefully you will reproduce the same 500-hPa map that is shown on the webpage linked above. If so, congratulations, you’re well on your way!

• As you go, you might want to save your notebook by clicking the ‘save’ button in the upper left.

• The first time you run things, there may be some map files that need to be downloaded, which takes a little longer and gives a warning. This is normal and won’t happen again after those files have been downloaded. You might also get some warning messages, but as long as the map plots and looks right, you can ignore them.

• Now spend a little time experimenting with the code in this notebook, to get a sense for some of the options (for example, try changing the contour intervals, or the map boundaries, or the vertical level shown, etc.)

• You may want to learn more about notebooks and how they work: a good resource is the Unidata training at https://unidata.github.io/python-training/workshop/Jupyter_Notebooks/jupyter-notebooks-introduction/

• When you’re done with JupyterLab, simply go under the File menu and select “Shut down”. Once this happens, you can close that browser window.

Lab 1 assignment

Now, we’re going to use some of these approaches to plot a surface weather map to analyze in multiple ways. For those of you with a lot of meteorological background, some of this may seem simple, but we want to make sure everyone’s on the same page before moving on to more complicated analyses.

Surface map

We’ll start by plotting a surface map from 0000 UTC 18 January 2025 over eastern North America. I’ve provided an example notebook that you can use for this (adapted from a MetPy example; you can right-click and save from this link): https://github.com/russ-schumacher/ats641_spring2025/blob/main/lab1/Station_Plot.ipynb. You’ll also need the file with all of the surface observations (in METAR format, which MetPy nicely decodes), here: https://github.com/russ-schumacher/ats641_spring2025/blob/main/lab1/metar_20250118_0000.txt.

Go through the steps in the notebook to get your surface map. Print it out, or if you have a tablet with a pencil you could do the analysis that way too. (If you don’t have easy access to a color printer, let Russ or Jacob know.)

Analyze the map given the guidelines below. For this analysis, focus on the synoptic-scale features by keeping your contours fairly smooth.

• Draw isobars in solid black lines at 4 hPa intervals.
• Draw isotherms in red (or another color) lines at 5°C intervals.
• Indicate the locations of any high and low pressure centers with an H or L, respectively.
• Analyze any fronts that are present.

In a couple sentences, describe the key features of the weather pattern that were revealed from your surface analysis.

MetPy automated analysis

Manual analysis is a valuable way to get a real “feel” for the data in a given weather situation, but it also can be time consuming. A wide variety of methods for automating the analysis of weather data have been developed over the years, with varying complexity. MetPy has some of these methods built in. We’ll use one of those methods here to analyze the same data that you analyzed by hand above.

Another example notebook, modified from a MetPy example, is at https://github.com/russ-schumacher/ats641_spring2025/blob/main/lab1/METAR_data_interpolation_dist.ipynb This will plot the same surface map as before, but will also analyze the pressure and temperature and plot them on the map. Go through this notebook to get the map.

Are there any noticeable/relevant/interesting differences between your hand analysis and this analysis?

You might not like how this map looks, so spend some time making a nicer-looking version of this analysis. The MetPy notebook that my example was based on (linked near the top) provides some ideas, but feel free to experiment with your own style.

GFS analysis

Finally, we can see how the analysis of solely surface observations compares to an operational forecast system with sophisticated data assimilation. Here, we’ll use NOAA’s GFS (but you could replicate this with other modeling systems if you want). In the same notebook above, steps are included to acquire the needed data using a magical tool called Herbie. Plot surface map(s) from the GFS following similar steps to what you used above.

Are there any noticeable/relevant/interesting differences between your hand analysis and the GFS analysis? What about between the interpolation method and the GFS? Discuss the strengths and weaknesses of each method, at least as far as you can determine from this single case.

When you turn in your write-up, include your hand-analyzed map, the initial map with the automated analysis, and your “final product” maps (hard copy and/or electronic versions are fine).