AI in Crop Yield Prediction Quiz

Answers and explanations appear as you take the quiz.

1. What does the NDVI vegetation index primarily indicate?
   • Soil pH level
   • The price of grain
   • Crop greenness and vigor from satellite light reflectance
   • The number of farm workers
2. Why do early-season yield forecasts carry more uncertainty than late-season ones?
   • Much of the weather and crop development that affects yield hasn't happened yet
   • Satellites turn off early in the season
   • Farmers hide their fields early
   • NDVI only works in winter
3. What does held-out-year validation help assess in a crop-yield forecast study?
   • A guarantee that no future information can leak into the inputs
   • Only whether the model's filename contains the year
   • Proof that every field shares the same yield
   • Generalization to another year's conditions, with forecast-time feature availability still needing control
4. Which data source provides the multispectral imagery commonly used in these models?
   • Home weather stations only
   • Sentinel-2 and Landsat satellites
   • Social media posts
   • Grocery store receipts
5. What does the cited US Corn Belt hybrid yield study add to machine-learning predictors?
   • Variables supplied by a process-based crop simulation model such as APSIM
   • Removal of every weather and soil input
   • A guarantee that the model needs no regional validation
   • Only random labels unrelated to crop development
6. Why should crop-yield evaluations examine unusual heat or drought years separately?
   • Extreme conditions always make prediction easier
   • A high NDVI guarantees an unchanged yield under every stress
   • Training mainly on ordinary conditions may not represent those stress responses adequately
   • Every weather year is identical once a crop is named
7. If a vegetation observation has near-infrared reflectance 0.6 and red reflectance 0.2, what is its NDVI?
   • 0.8, by adding the two values only
   • 0.5, using (0.6 minus 0.2) divided by (0.6 plus 0.2)
   • 0.4, by subtracting without normalization
   • 3.0, by dividing near-infrared by red only
8. What does validation against county-average yield labels establish about an individual-field forecast?
   • It guarantees every field has the county's average yield
   • It converts a regional label into a direct harvest measurement for each field
   • It removes all need to know the forecast's spatial unit
   • It does not by itself establish field-level accuracy; the intended scale needs its own evidence