Gemini Optimized

Best Gemini prompts for Atmospheric and Space Scientists

A specialized toolkit of advanced AI prompts designed specifically for Atmospheric and Space Scientists.

Professional Context

Balancing the urgency of analyzing satellite data to predict severe weather events with the need to refine climate models, Atmospheric and Space Scientists face a daily tension between immediate public safety and long-term research goals. As they navigate this tightrope, they must leverage cutting-edge tools and workflows to maximize the impact of their work.

💡 Expert Advice & Considerations

Don't waste time trying to use Gemini for original research; instead, focus on using it to automate tedious data cleaning and interpretation tasks, freeing you up to focus on higher-level analysis and decision-making.

Advanced Prompt Library

4 Expert Prompts
1

Interpreting Spectral Data from Space-Based Observations

Terminal

I have a dataset of spectral readings from the GOES-16 satellite, covering a 24-hour period over the eastern Pacific. The data includes radiance values at 16 different spectral bands, as well as corresponding latitude and longitude coordinates. Using Google Earth Engine, write a script to: 1) import the dataset and convert it to a usable format, 2) apply a spectral unmixing algorithm to identify the presence of different atmospheric constituents, and 3) visualize the results as a time-series animation, with a separate layer for each constituent. Be sure to include a legend and scale bar in the visualization.

✏️ Customization:Replace the dataset and spectral bands with your own data and parameters.
2

Analyzing Climate Model Output with Google Cloud AI Platform

Terminal

I've run a climate model simulation using the CESM2 framework, generating a large dataset of output files in netCDF format. Using Google Cloud AI Platform, write a pipeline to: 1) ingest the output files into a Cloud Storage bucket, 2) apply a set of predefined data processing scripts to extract relevant metrics (e.g. global mean temperature, sea level pressure), and 3) train a machine learning model to predict future climate trends based on historical patterns in the data. Be sure to include hyperparameter tuning and model evaluation steps in the pipeline.

✏️ Customization:Modify the pipeline to accommodate your specific climate model and output variables.
3

Generating Reports on Severe Weather Events from Radar Data

Terminal

I have access to a database of radar reflectivity data from the NEXRAD network, covering a recent severe weather event in the central United States. Using Google BigQuery, write a query to: 1) extract the relevant data for the event, including reflectivity values, storm motion, and demographic information for affected areas, 2) apply a set of predefined thresholds to identify areas of significant damage or loss, and 3) generate a report summarizing the event, including maps, tables, and statistics on the impacted regions. Be sure to include a section on recommendations for future event preparedness and response.

✏️ Customization:Update the query to reflect the specific event and data parameters you're working with.
4

Validating Satellite Precipitation Estimates against Ground Truth Data

Terminal

I've obtained a dataset of satellite-based precipitation estimates from the GPM mission, covering a 6-month period over the continental United States. I also have access to a corresponding dataset of ground truth precipitation measurements from the NOAA gauge network. Using Google Earth Engine, write a script to: 1) import and align the satellite and ground truth datasets, 2) apply a set of statistical metrics (e.g. mean absolute error, correlation coefficient) to evaluate the accuracy of the satellite estimates, and 3) visualize the results as a series of maps and scatter plots, highlighting areas of agreement and disagreement between the two datasets. Be sure to include a discussion of the implications for hydrological modeling and prediction.

✏️ Customization:Replace the datasets and metrics with your own data and parameters.