Module 03 · Adapt · Be An Earth System Trend Detective!
Try a crop rotation on your field and see the trade-offs before you plant.
Interactive and fully static: everything runs in your browser from precomputed data. Use the "Experiment briefing" button inside the tool for a guided walkthrough.
A farmer choosing next year's rotation has to weigh soil, water, erosion and a changing climate, but the satellite evidence for their own field is scattered across agencies and formats.
Farmers face shifting rainfall, heat, frost and declining soil health, and the rotation they pick this year shapes the soil for years.
Field Shift combines NASA Earth observations with local soil, crop and farmer priorities for Story County, Iowa, at the scale of a 1.2 km tile.
The data story · Adapt
Five beats: the raw signal, what changed, how sure we are, where it gets subtle, and who acts on it.
01 · The signal
Story County, Iowa, is mostly corn and soybeans. USDA's Cropland Data Layer gives each 30 m field a crop label every year since 2009, and NASA's Harmonized Landsat and Sentinel-2 (HLS) shows how green each field is in mid-summer.
Field Shift learns from that history, plus NASA terrain, climate and NDVI, to predict next season's crop.
02 · What changed
NASA POWER daily temperatures for central Iowa, 1981–2025, show the frost-free season growing by +0.43 days a year. Most of it comes from autumn: the first frost now arrives later (p < 0.001), while the last spring frost hasn't moved (p = 0.48). Corn growing degree days rise too (p < 0.001).
03 · How sure we are
The 95% confidence interval for the season trend is +0.12 to +0.67 days a year, entirely above zero. p = 0.006: a pattern this strong would turn up by chance about 6 times in 1,000 with no real trend. The Story County cell agrees (p = 0.002).
Water is the other half of the decision. GRACE, GLDAS, SMAP, GPM IMERG and NASA POWER all find no significant trend in Iowa's water, so a longer season isn't coming with less water.
Is Iowa's water changing?
Five records agree: steadyIs Iowa's growing season getting longer?
One source, two scales: longer04 · The nuance
Mid-summer NDVI is near saturation over both corn and soybeans, so HLS shows field-by-field contrast but can't separate crop types. Crop labels come from the Cropland Data Layer, not from HLS. SMAP soil moisture covers only 11 seasons, too short for a strong test.
The crop model is honest about its limits: hay F1 is 0.33, the rare 'other crop' class is 0, and temperature scaling made calibration worse, so the tool keeps the raw probabilities.
05 · Who acts on it
A farmer or agronomist decides what to plant next year and how to rotate over four. A later first frost with steady water widens the window after harvest, which favours longer-season hybrids and cover crops: options the rotation explorer lets them test.
It ranks 36 non-dominated four-year plans by the farmer's own weights on five indicators: water demand, erosion risk, nitrogen-fixing share, heat and frost exposure, and soil-moisture stress.
Launch the dashboardWhat this means
≈19 dayslonger frost-free season since 1981Central Iowa, NASA POWER, +0.43 days a year (p = 0.006).
With no significant trend in Iowa's water across five records.Is Iowa's water changing?
Five records agree: steadyIs Iowa's growing season getting longer?
One source, two scales: longerWhat this means
≈19 dayslonger frost-free season since 1981Central Iowa, NASA POWER, +0.43 days a year (p = 0.006).
With no significant trend in Iowa's water across five records.Who uses it
The path through the dashboard, from the situation they start in to the decision it changes.
Story County, Iowa
They run corn-soy and keep hearing the climate is changing, but don't have their field's history, their priorities and a tested season-length statement in one place.
Click a 1.2 km tile on the 2025 crop map, or pick the county-typical tile.
Move the priority sliders: water demand, erosion risk, nitrogen-fixing share, heat/frost exposure, soil-moisture stress.
Compare current corn-soy with the ranked four-year plans, the trade-off table and the Pareto front.
Read the trend layer: longer, warmer season; no significant water trend.
Pick a rotation that uses the longer autumn window (a cover crop after harvest, or a longer-season hybrid) without budgeting for a wetter or drier summer.
A shortlist scored to their own weights, a tested climate statement, and a reminder to check it with a local agronomist.
An illustrative journey built from what the dashboard does today, not a field-tested outcome.
Launch the dashboardAdvising several Iowa farms
They won't recommend a tool they can't audit; they need to know where the model is weak and which climate claims hold up.
Plan a rotation on a tile, as the farmer would.
Open "How we scored it" and "Model honesty": baselines, ablation, the calibration miss, the 2025 disagreement map.
Read the HLS and SMAP caveats and the crop and water trends.
Use the tool to talk through season length and trade-offs, and override it on hay, rare crop classes and any soil-moisture trend claim.
They can point at a tile and say what is tested, what is a weighted score, and what still needs a field walk.
An illustrative journey built from what the dashboard does today, not a field-tested outcome.
Launch the dashboardNext-season crop accuracy on 66,666 held-out pixels (2024–2025).
Macro-F1 across six crop classes. Adding NASA terrain, soil, climate and NDVI to crop history lifts it from 0.4582.
Pixel accuracy of the full-county 2025 map.
Erosion-risk index of a cover-crop rotation vs the current corn-soy rotation (county-typical field).
Median growing-season rainfall used in the climate panel.
Honest miss: temperature scaling made calibration worse (ECE), so the tool keeps the raw probabilities. Hay F1 is 0.33 and the rare 'other crop' class is 0.
Held-out test data, compared against baselines. One seed per run, not a published benchmark.
Visuals
Earthdata upgrade · NASA HLS
We added NASA's Harmonized Landsat and Sentinel-2 (HLS v2.0, 30 m), downloaded with an Earthdata login, for Story County, Iowa (tile T15TVG). Comparing clear mid-summer scenes on 1 August 2024 (Sentinel-2) and 31 July 2025 (Landsat) over 866 km² of vegetated land, the median NDVI change is −0.01; 42 km² got greener by more than 0.1 and 22 km² got less green by more than 0.1.
Two single dates from two sensors in two years. Mid-summer NDVI is near saturation over both corn and soybeans, so HLS here can't separate crop types: it shows field-by-field contrast, not a crop map or a trend. The 2024 Sentinel-2 scene covers only about 60% of the county. We also processed SMAP enhanced L3 surface soil moisture (9 km, Earthdata login, 1,602 daily files): May–September means of 0.16 to 0.22 m³/m³ in Story County over 2015–2025, with no significant trend (p = 0.47; central Iowa p = 0.21). Eleven seasons is a short record.
Trend layer · Crops
Rotation choices depend on season length, heat and soil water. We built seven growing-season indicators from 45 years (1981–2025) of NASA POWER daily data for central Iowa and the Story County cell where Field Shift runs, and tested each for a trend.
Mann-Kendall (α = 0.05) and Sen's slope with 95% CI, one value per year. Hover or tap the chart for yearly values.
Where it is changing
The approach
USDA Cropland Data Layer 2009–2025 at 30 m, grouped into 923 analysis tiles of 1.2 km.
NASADEM slope, NASA POWER climate (growing-degree days, frost, heat, rainfall, root-zone wetness), MODIS NDVI bare-soil months and SoilGrids texture and carbon.
LightGBM predicts each pixel's next crop from five years of history plus terrain, soil, climate and NDVI. Trained to 2021, tuned on 2022–2023, tested on 2024–2025.
Five indicators per rotation: water demand (FAO-56 crop coefficients), erosion risk, nitrogen-fixing share, heat/frost exposure and soil-moisture stress.
Three reference rotations plus an NSGA-II search (48 individuals, 25 generations) that finds 36 non-dominated four-year plans. Sliders re-rank them by the farmer's own priorities.
Limits we know about
| Dataset | How it's used | Access | Link |
|---|---|---|---|
| NASA POWER | Daily climate 2000–2024: temperature, rainfall, root-zone wetness | Public, no login | source |
| NASADEM | Elevation and slope (NASA JPL / LP DAAC) | Public, no login | source |
| MODIS MOD13Q1 NDVI | 16-day vegetation index; bare-soil months | Public, no login | source |
| USDA NASS Cropland Data Layer | 30 m crop maps 2009–2025 (public domain) | Public, no login | source |
| ISRIC SoilGrids | Clay, sand, organic carbon, pH, bulk density (CC-BY-4.0) | Public, no login | source |
| NASA POWER daily T2M_MIN, T2M_MAX, GWETROOT, PRECTOTCORR (trend layer) | 1981–2025 growing-season indicators, central Iowa (35 cells) and Story County; Mann-Kendall + Sen's slope | Public, no login | source |
| GRACE/GRACE-FO JPL mascon RL06.3 CRI v04 (trend layer) | Monthly total water storage 2002–2026, land cells over central Iowa; deseasonalised annual means | Earthdata login | source |
| NASA HLS v2.0 HLSL30 + HLSS30 (module upgrade) | 30 m NDVI, Story County, two mid-summer scenes (1 Aug 2024, 31 Jul 2025) | Earthdata login | source |
| NASA SMAP L3 enhanced soil moisture SPL3SMP_E (module upgrade) | 9 km surface soil moisture, Story County and central Iowa, May–Sep means 2015–2025; Mann-Kendall + Sen's slope | Earthdata login | source |
| GLDAS Noah 2.1 monthly GLDAS_NOAH025_M (trend layer) | Soil moisture at four depths to 1 m, central Iowa, May–Sep means 2000–2025 | Earthdata login | source |
| GPM IMERG V07 monthly GPM_3IMERGM (trend layer) | Satellite rainfall, 0.1°, central Iowa box, annual totals 1998–2024 | Earthdata login | source |
For the judges
Helps farmers and advisers compare rotations on soil, water and climate risk, field by field, before committing a season.
Combines a next-season crop model with a multi-objective search, so farmers explore a trade-off frontier instead of getting one 'best' answer.
Temporal split, three baselines, ablation by data layer, calibration checks and per-class scores, including the classes it gets wrong.
Trend Detective: 45 years of NASA POWER show central Iowa's frost-free season and growing degree days rising significantly, which changes the rotation question. NASA POWER, NASADEM and MODIS NDVI drive the indicators and the model.
Click a tile, move five priority sliders and watch the ranked rotations update, with a plain-language note on each plan's benefits and risks.
First designed around the related Space Apps 2026 brief Field Shift: Adapting Farms with NASA Data; here it is the adapt module of one system in Be An Earth System Trend Detective!