Pick a question. TerraTrend will choose the variable and region, fetch
NASA data, and test whether the change is statistically significant.
You can change anything afterwards.
Or close this and choose your own variable and region.
WORLD / CONTINENTS
Select an area—
Last completed result
—run an analysis
—·—
Choose a variable and area, then analyze.
DEFINE YOUR AREA
Select primary area A
Choose a geographic level above, then click an area on the globe.
0 points
AREA COVERED · APPROX.
Click to place or move the centre. Drag to navigate; scroll to zoom.
LEFT: YEAR ADrag handle left/right to reveal change across coordinatesYEAR B: RIGHT
Physical Corroboration: NASA GIBS imagery is orthorectified and surface-reflectance calibrated. Comparing identical calendar dates across decades isolates multiannual environmental trends (such as vegetation expansion, aridification, or ice margin retreat) from regular seasonal cycles.
TERRATREND · SCIENTIFIC OBSERVATORY DEEP DIVE
HIGH-RESOLUTION TREND ANALYSIS
REGION · SAMPLES · ALPHA: 0.05
SERIES:
ZOOM AXIS:
ZOOM: 1.0xScroll / wheel to zoom · Swipe / drag to pan · Double-click to reset
SPATIAL GRID CONSENSUSNATIVE CELL DETAIL · ACTUAL COORDINATESNATIVE CELLS▸
Hover a native cell to inspect its coordinate-specific trend.
EARTH SYSTEM COUPLING & MULTI-SPHERE FEEDBACK0 DETECTED▸
REGIONAL CONTRAST DEEP DIVE (REGION A VS REGION B)CONTRAST ACTIVE▸
GLOSSARY · PLAIN ENGLISH
WHAT THESE VARIABLES MEAN
Every measurement explained without jargon — for students, journalists and Earth observers
TREND DISCOVERY · SUPPORTED MEASUREMENTS
REGION
—
Ranked by significance, then by trend coherence (|τ|). Uses the same native-cell
grid as the full analysis, so p-values match when you open a row.
Downloaded subsets are reused; opening a measurement may still take calculation time.
Cells shown as significant / total. Significance is per measurement; discovery is exploratory.
EARTH SYSTEM IMPACT LEDGER · PLANETARY TELECONNECTIONS
GLOBAL RIPPLE EFFECTS & TELECONNECTIONS
Ranked planetary cross-correlations and physical cascading mechanisms
ORIGIN REGION—
ORIGIN VARIABLE—
SIGNIFICANT RIPPLES—
STRONGEST COUPLING—
Statistical cross-correlation with Student's t significance (p < 0.05) across NASA POWER multidecadal grid points.
EARTH SYSTEM NETWORK · COMPUTED INTERCONNECTIONS
INTERCONNECTIONS
Select an area and measurements to investigate statistical relationships and possible temporal leads.
—
Search details and anomaly summary
WITHIN-REGION COUPLING
Relationships and possible temporal leads between measurements in this area.
POSITIVEINVERSEARROW = COMPUTED LEAD
REMOTE RELATIONSHIPS
Remote regions whose variables respond to this driver's anomalies.
DRIVER ANOMALY EPISODES
When this variable departed from its own seasonal cycle and trend.
NO SIGNIFICANT LINKS DETECTED
The driver's anomalies do not explain any remote variable beyond what chance and the shared seasonal cycle already account for. That is a real scientific result, not a failure — it means this signal is localized.
Choose an event. Observation checks start only after you review a request.
OPEN DATA · VISIBLE METHODS
Know what is behind the map.
Geographic boundaries define the area. Scientific datasets define what we can say about it.
PROCESSING SETTINGS
Automatic processing, visible sources.
Native MERRA-2: eight measurements clipped to the selected boundary. Giovanni: monthly temperature and humidity at 2 m and 10 m for rectangles, local relationships and batches of preset boxes. AppEEARS: monthly polygon NDVI/EVI summaries. Hydrology: monthly GLDAS native-cell histories. NOAA: fixed-cohort station observations, not whole-area averages. Linked datasets: combine source-qualified measurements over the same area and calendar. Each source keeps its own resolution, weighting and quality limits; values need not match. No source silently falls back to a larger raw-data download.
TerraTrend automatically discovers monthly NASA MERRA-2 observations through CMR and downloads native-grid subsets from GES DISC, the archive behind the challenge’s Giovanni resource. Each result retains the dataset citation, units, native cell spacing, coverage and geographic boundary version.
NASA POWER is another NASA service. Its existing point-cache files remain available for continuity. A live POWER option still needs verified mappings and native-grid coverage checks; this map workflow makes no new POWER requests.
02 / THE MEASUREMENT
Whole areas. Honest coverage.
In native mode, grid cells are clipped to the chosen boundary and weighted by their overlap and latitude. At least 80% area support and complete calendar years are required. Giovanni mode uses NASA’s rectangle averages and does not report monthly spatial support. Small regions can share a coarse source cell: a sharp boundary does not add scientific resolution.
Theil–Sen measures the trend; Mann–Kendall tests significance. Monthly seasonality, confidence intervals and autocorrelation checks are available in Analysis. Paired calendar-aligned series support A/B comparisons.
Remote links represent statistical associations and possible lead relationships, not established physical causes or measured wind trajectories. Anomaly relationships do not by themselves explain a long-term trend.
03 / GEOGRAPHIC REFERENCE
One boundary, every view.
Global country boundaries use Natural Earth 1:50m, including the disconnected parts present at that resolution. Continental and subcontinental layers follow UN country membership; transcontinental countries remain whole in this scheme.
All 16 Ghana regions are bundled. Other countries load first-level administrative boundaries on demand through geoBoundaries gbOpen when available. Reference boundaries are separate from scientific measurements.
Choose an area and period, then Analyze. TerraTrend fetches the published monthly grids, clips them to your actual boundary and caches them for reuse. No file uploads are needed.
Checking the NASA connection…
Start the updated server
In the terminal running TerraTrend, press Ctrl+C.
From the project folder, run ./start.sh 8517.
Reload this page, then connect Earthdata below.
Your selected areas stay prepared while you view these instructions. A page reload starts a new investigation.
VERIFIED AUTOMATIC MEASUREMENTS
Air temperature · Surface temperature · Specific humidity · Surface pressure · Root-zone soil moisture · Surface soil moisture · Land evaporation · Precipitation
Other catalogue variables still need verified source mappings. Monthly wind components cannot substitute for mean wind speed; solar radiation products cannot substitute for direct normal irradiance. Missing observations remain missing.
First downloads may take several minutes for large areas or long periods; progress and cancellation are available. Source granule IDs, collection version, observed dates and area coverage accompany each result.
05 / MONTHLY VEGETATION
Connect AppEEARS.
Use your NASA Earthdata username and password to obtain AppEEARS’ separate service token. The password is used only for NASA’s login request and is never saved. With Remember enabled, only the issued token is saved privately on this computer. NASA normally expires it after about 48 hours.
AppEEARS connection not checked yet.
If NASA needs an authorization step, sign in to AppEEARS ↗ and accept its account prompts, then connect here. Choose monthly NDVI/EVI for polygon averages. NASA may create raster outputs remotely; TerraTrend retrieves only the statistics CSV. No additional pixel-quality filter or monthly area-coverage audit is claimed.
06 / REGIONS & LINKED EVIDENCE
Focus, cache, investigate.
Use ◇ Presets on the map to return to Ghana Savannah, Congo Basin and the original focus boxes. These are curated rectangles, not official ecosystem boundaries. Analyze a box to cache its observations; repeat matching requests can reuse them. Selection alone downloads nothing.
Choose Linked datasets and select two or three measurements—for example root-zone soil moisture and vegetation greenness—then open Interconnections. Local screening is the default. Enable the separate remote search only when you want additional areas included.
NOAA monthly station observations use a public access API and require no additional key. Station summaries are independent supporting observations, not substitutes for whole-area averages. Measurements from different sources retain their own definitions, resolution and quality limitations.