Challenge 11 - Chronicles of Drought: Contextualising Earth's Water Story

[RubenRT7]

Challenge 11 - Chronicles of Drought: Contextualising Earth's Water Story

Stream 1 - Data Visualization and visual narratives for Earth Sciences applications

Goal

"DroughtChronicles" is a project that seeks to contextualize the understanding of drought events on Earth. We propose the development of a series of Jupyter notebooks that will vividly illustrate our planet's drought stories in an accessible and engaging manner.

Mentors and skills

• Mentors: Christopher Barnard, Francesca Di Giuseppe, Jessica Keune, Fredrik Wetterhall (all ECMWF)
• Skills required:
  • Python language proficiency
  • Good understanding of data analysis
  • Visualisation literacy

---

Challenge description

Droughts are prolonged periods of abnormally low precipitation that result in water shortages, adversely impacting ecosystems, agriculture, and communities. Characterized by an insufficient water supply to meet the demands of various sectors, droughts can lead to crop failures, water scarcity, ecological imbalances and increase the risk of wildfires, posing significant challenges to both rural and urban areas. Climate change is contributing to an escalation in the frequency and severity of droughts, amplifying their impact worldwide.

To monitor the changing patterns of droughts, ECMWF has developed a drought library, DRYFALL, enabling the calculation of a significant array of indicators. Building on this resource, "DroughtChronicles" is a project that seeks to contextualize the understanding of drought events on Earth. We propose the development of a series of Jupyter notebooks that will vividly illustrate our planet's drought stories in an accessible and engaging manner.

These notebooks will retrieve and process data from the Common Climate Data Store (CDS), transforming raw information into expressive visual narratives. We will focus on calculating climate anomalies, spatial extent and severity of events and generating visualizations that echo the vibrant storytelling found in journalistic chronicles.

The resulting Jupyter notebooks will not only offer meaningful climate insights but also serve as a comprehensive training resource. Through DroughtChronicles, we aim to provide a more visual, comprehensible, and reproducible approach to improve our understanding of the impact of drought on our society.

We aim to generate at least 3 workflows able to answer the following questions

1. What regions of the world are or have been in drought in a certain year and how severe was the drought?
2. How anomalous was that drought?
3. What were the drivers of the drought? Was it ‘just’ a lack of precipitation? Did extreme temperatures contribute to the drought occurrence and severity?

For any of these notebooks we expect the following steps to be followed:

• Define the Notebook Purpose and Structure:
  Clearly outline the purpose of the notebook. Decide on the structure, including sections, sub-sections, and a logical flow that will help readers follow the content easily.

• Setup and Environment:
  Begin the notebook with instructions on setting up the Python environment. Include information on installing any necessary libraries or dependencies. Use code cells to demonstrate the installation process.

• Introduction and Overview:
  Provide a brief overview of the key concepts or tasks that will be addressed. Include any background information that is essential for understanding the content.

• Data Preparation and Loading:
  Create a section for data preparation and loading. Use code cells to demonstrate how to load datasets, handle missing values, and perform any preprocessing steps necessary for the analysis.

• Main Content and Code Implementation:
  Divide the notebook into sections corresponding to the main content Clearly label and comment on each code cell to guide readers through the steps. Encourage readers to run code cells interactively.

• Visualizations and Results:
  Include sections for data visualizations and displaying results. Use Matplotlib, Seaborn, or other relevant libraries to create plots and charts that complement the analysis. Add markdown cells to explain the significance of visualizations and findings.

• Conclusion and Summary:
  Conclude the notebook with a summary of key findings, takeaways, or lessons learned.

The successful candidate will work with a team of meteorological applications experts who have developed data analysis and an initial version of the workflow. It will be a great opportunity to collaborate with senior scientists and gain an understanding of how quality data can be transformed into a compelling story to be told.

[aalonsop]

Hello!
This project seems very interesting. Is it possible to have a link to the dryfall library? Or is it internal only?
thanks

[fdg10371]

Thank you for your interest in this challenge. The DRYFALL library is still under development and is not publicly available yet. However, the challenge does not require the development of the library but rather the utilization of the data produced to generate information content in a manner accessible to a wider community of users. We can make a sample dataset available to showcase the type of level zero information upon which we require development

[aalonsop]

Yes that would be great! It will help a lot with the proposal.

[enyfeo]

I will make something available early next week and add it to the challenge information.

[enyfeo]

Example SPEI12.zip

The Standardized Precipitation Evapotranspiration Index (SPEI, Vicente-Serrano et al., 2010) is a commonly used meteorological drought index. The SPEI measures the water deficit at the land surface that accumulated over a certain time windows, usually months, and evaluates the deficit with respect to a reference period. SPEI values are in units of standard deviation from the standardised mean, i.e., negative values indicate drier-than-usual periods while positive values correspond to wetter-than-usual periods. The magnitude of the SPEI is an indicator of the severity of event. The following range of values is typically used to identify the severity of the event:

SPEI > 2.0: extremely wet
1.5 < SPEI <= 2.0: severely wet
1.0 < SPEI <= 1.5: moderately wet
0 < SPEI <= 1.0: near-normal / mildly wet
–1.0 < SPEI <= 0: near-normal / mildly dry
–1.5 < SPEI <= –1.0: moderately dry
–2.0 < SPEI <= –1.5: severely dry
SPEI < –2.0: extremely dry

Values of the SPEI between -1 and 1 are also often considered near-normal.

The SPEI is typically computed over a range of time windows from 1 over 3 and 6 to 12 months or more. The time window considered is indicative of the potential impact of meteorological drought, which is often the primary driver of drought.

Here, we provide examples of the SPEI-12:
A global snapshot of the SPEI-12 for November 2023, indicating the water deficit at the land surface that accumulated between November 2022 and 2023.
A monthly timeseries of the SPEI-12 from 1991–2023 for all grid points over Mozambique and surrounding.
To calculate the SPEI, the WMO suggested reference period of 1991–2020 is used and a general logistic distribution is fitted to the the difference (precipitation – potential evaporation) from ERA5 at each grid point and each month. The parameters of that fitted probability density function are then used to find the cumulative probability of the observed deficit for the required month and time window. Finally, this cumulative probability is then transformed to the standardised normal distribution with mean 0 and variance 1, providing the final value of the SPEI. The original values of the SPEI are provided and drought occurrence and severity can be identified using the thresholds defined above. High values in either direction (values smaller than –3 or larger than 3) should not be interpreted in terms of severity as the tails of the distribution are likely not well represented. The data provided here is not masked by significance yet.

[liamcassidy3]

Hello Chris @enyfeo ,

I hope all is well with you.

My colleagues and I were about to send in our proposal, but it looks like the "submit your proposal" link is down on the code for earth website (https://codeforearth.ecmwf.int/). Is there an alternative way we should plan to submit our proposal before the deadline expires today?

Any help is greatly appreciated!

-Liam

[trakasa]

@liamcassidy3
thanks for the information and letting us know about your issues with the submission form. This should have been sorted out and you should be able to submit your proposal.

Please let us know, if you still have any issues.

Apologies for the inconvenience and thanks for your patience!

Bye, Athina