Scientific research workspace

Design an investigation

Select where to look, what to measure and over which period. TerraSignal retrieves the observations, validates them and runs the statistical tests — you will see the analysis in the Trend Lab.

NASA POWERNASA CMRNASA GIBS
Step 1 · Where

Study location

Search a place name, type coordinates, or click anywhere on the map.

Nepal Himalayas28.000°, 84.900°· source: TerraSignal gazetteer

The dashed box is the native 0.5° × 0.625° MERRA-2 cell that NASA POWER samples. TerraSignal never draws a finer surface than the data supports.

Step 2 · What

Environmental variables

Up to six parameters can be tested in one run.

1/6 selected · multiple variables are tested together with Benjamini-Hochberg FDR control.

Step 3 · When

Observation period

POWER daily record begins 1981-01-01

Latest available data lags real time by a few days

A trend needs a decade or more to be meaningful; Mann-Kendall is not reported below 10 annual values.

≈ 13,148 daily records per variable will be requested.

Step 4 · Method

Analysis settings

Daily retrieval is aggregated by TerraSignal; monthly uses POWER's own monthly product.

Determines the parameter set POWER returns for the surface meteorology.

0.05 is the scientific convention, not a law of nature. Lower α is stricter; results with p slightly above α are not evidence of absence.

Step 5 · Run

Research question

Optional, but it is what the generated report will answer.

Investigation summary
Location
Nepal Himalayas (28.000°, 84.900°)
Variables
T2M
Period
1991-01-01 → 2026-12-31
Method
daily product · α = 0.05 · RE
Provider
NASA POWER (Renewable Energy community)
Engine
checking…
Live retrieval from NASA POWER · typically 1–4 s
What will happen

Pipeline stages executed by this run

  1. 01Dataset discoveryThe verified POWER product is matched to your variables and resolution.
  2. 02Metadata retrievalProvider, product, API version and model sources are recorded.
  3. 03Data collectionDaily values are requested for the exact grid cell containing your point.
  4. 04Quality validationMissing values, fill values, implausible values and irregular intervals are flagged.
  5. 05PreprocessingUnits standardised, timestamps normalised, aggregation rules recorded.
  6. 06Trend detectionMann-Kendall (autocorrelation-corrected) + Sen's slope + OLS cross-check.
  7. 07Anomaly & change pointsz-score, modified z-score, Isolation Forest and Pettitt segmentation.
  8. 08Scientific validationFDR control, effect size, completeness and confidence grading.