Skip to content
All challenges

Be An Earth System Trend Detective!

AdvancedSubjects:Earth ScienceSoftware

Your challenge is to create a tool that can select variables measured by Earth science missions or produced by Earth science models, visualize how they change over time, and determine what is changing, where it is changing, how much it is changing, and if these changes are what scientists call “significant.”

Illustration of Earth with four quadrants showing different global data layers, including weather, precipitation, land surface, and atmospheric patterns.
Image: NASA Space Apps Challenge

Summary

Change is always occurring in the Earth’s interconnected environmental system, and when it runs in a consistent direction – rising or falling, increasing or decreasing, thickening or thinning – measured variables reflect it. But a measured Earth science data variable can trend one way in one region and the opposite way in another, even when the same process drives both. The multitude of Earth science data variables makes it difficult to select variables that could indicate change, and to then perform an analysis on the selected variables. Your challenge is to create a tool that can select variables measured by Earth science missions or produced by Earth science models, visualize how they change over time, and determine what is changing, where it is changing, how much it is changing, and if these changes are what scientists call “significant.”

Plays from YouTube

Background

Instruments on satellites observe many different aspects of the Earth system: on land, in the atmosphere, and in the oceans. Global models also use satellite instrument data and other kinds of data to examine and analyze the Earth system. One important reason for doing this is to observe and characterize changes in the Earth system. Changes, especially those moving in a particular direction over time – a trend – indicate both the processes causing such changes and the processes that will respond to them.

One complicating factor in this work is the wide variety of variables that can be studied. Another is that changes in a particular variable can move in one direction in one region and in the opposite direction in another, even when both trends are caused by the same process! So there are two important parts to this problem: finding variables that are trending, from simple ones like temperature and rainfall to more complex ones like evaporation or streamflow; and analyzing trends in particular regions of the land, ocean, or even levels of the atmosphere.

Instruments such as the Moderate Resolution Imaging Spectroradiometer (MODIS), the Ozone Monitoring Instrument (OMI), the Atmospheric Infrared Sounder (AIRS), and the Global Precipitation Measurement (GPM) provide analyzable variables. Global models – notably the Modern Era Retrospective analysis for Research and Applications, Version 2 (MERRA-2) and the Global Land Data Assimilation System (GLDAS) – provide many more. Finding, analyzing, and relating significant trends to the processes causing them is one of the most important uses of NASA Earth science data.

Objectives

A successful response to this challenge will bring together several important elements.

The first is to identify Earth science variables of interest that can be investigated for potential trends. As the Background notes, NASA and other space agencies offer numerous variables, some familiar and some more specialized for use by particular branches of science.

The second is to acquire the data variables for a particular region, since most trends in the data will not be global – though some trends can be!

The third is to generate and visualize a time series for the variable and region of interest over a long enough period of time to allow statistical analysis.

The final element of the challenge is to perform a basic statistical analysis of that time series, indicating the rate of change over time, and to test the trend’s significance using methods such as a t-test, the Mann-Kendall test, or confidence intervals.

Data sites offer tools that generate time series, and statistical analysis packages provide basic significance testing. A response that does both the visualization and the analysis meets the full scope of this challenge.

Potential considerations

You may (but are not required to) consider the following:

Audience

  1. This challenge involves science data and statistical analysis, so the likely audience is high school and college students, along with members of the public who have some scientific background.
  2. Providing simple definitions of the variables would help these users considerably.
  3. The concept is to help the users become data detectives – finding variables of interest, then determining whether those variables show a trend and whether that trend is significant.

Approach

  1. Users could potentially respond to a short series of questions about what or where they would like to investigate, and then the answers to the questions would allow searching of related data variables. It would be highly useful to provide a couple of demonstration examples that the user could watch and would help them understand how the tool works.
  2. While not required, SI could be useful here to identify data with the potential to show trends in particular areas, drawing on either news articles and research papers or the characteristics of a region such as a desert, rain forest, tidal estuary, or the atmosphere above a large city.

Working with the data

  1. Earth science data are available at several different temporal resolutions – hourly, 3-hourly, daily, and monthly. For this challenge, daily and monthly data are likely the easiest to process.
  2. Finding a trend is interesting even if it doesn’t reach the level of significance. Sometimes all a trend needs to become significant is a longer period for the data to be acquired.
  3. Consider looking for publicly available tools for plotting hydrological data offered by outside organizations like the Consortium of Universities for the Advancement of Hydrologic Science, Inc. (CUAHSI).

Trend analysis with Earth science data can be complex. Trends can be non-linear, and some are not evident unless regular cycles, such as the seasonal cycle, are removed. More complex trend analyses are beyond the scope envisioned for this challenge; detecting and analyzing linear trends would be a great start.

See the 13 resources for this challenge

From the NASA Space Apps Challenge. The official page is the final word on this challenge.

Built in Athlone

The 2 projects Athlone teams submitted for this challenge at HackAthlone 2026, from the teams’ own Space Apps pages. Links open in a new tab.

    • 2026
    • Best Use of Open Data

    Erevna

    Be An Earth System Trend Detective!

    Earth Trend Globe is an interactive 3D globe that shows how Earth’s climate has changed since 1981. It also shows which of those changes are real. We used 45 years of NASA POWER data (MERRA-2, CERES and GEWEX SRB) for six variables: temperature, humidity, rainfall, sunlight, air pollution and ozone. The data covers 16,020 points around the world. Users can click anywhere, choose a region, or draw their own area. Each area gets a trend per decade and a significance test that allows for month-to-month memory in the data. If a change isn’t clear yet, the app says how many more years of data it would need. A “data detective” mode suggests places worth exploring. Live storms, floods and fires from NASA EONET are linked to the long-term trend at the same spot. A statistical outlook, tested on 2008 to 2025, shows what 2040 could look like for 184 areas. This matters because a single heatwave or flood is easy to mistake for a trend. Our globe separates real change from natural ups and downs, and it explains everything in plain language. It is honest about what the data does and doesn’t show.

    Read moreShow less

    Team: Jovan Tomy, Ryan Jyothis, Dhananjai Sudarshan, Síofra Dennehy, Brandon Shumba

    • 2026

    TerraLens

    by The Cosmic Cartographers

    Be An Earth System Trend Detective!

    TerraLens starts from a place and finds what is really changing there. Pick one of five regions (Ireland, Iberia, the Sahel, NW South Asia or the US Southwest), and TerraLens tests five NASA Earth-science variables over 1980–2025: air temperature, soil moisture, evapotranspiration, precipitation and aerosol, all from MERRA-2 via NASA GES DISC. It ranks what is changing, shows how fast with a 95% confidence interval, and gives each variable a plain verdict. We adjust every p-value with the Benjamini–Hochberg method to keep false discoveries. Opposite Trends puts two places side by side: US Southwest rainfall is falling while NW South Asia’s is rising. The Time Machine shows how the verdict changes with the years you choose, and how long a record would be needed to detect a trend that isn’t proven yet. This covers every step the challenge asks for: choosing variables, getting regional data, building long time series, and measuring both the rate and the significance of each trend. It matters because charting tools draw a trend line without saying whether it is real. TerraLens tells students and the public what is changing, how sure we can be, and when the honest answer is “not proven yet”.

    Read moreShow less

    Team: Bhargav Lalit Chaudhari, Parth Soni, Kevin Henry, Nandini Hazarika