Dashboard / Problems / landsat-composite

Landsat 8 Annual Median Composite (Yellowstone, 2022)

temporal-analysis intermediate USGS Landsat 8 Collection 2 Level 2

Ground Truth

0.2
reflectance
15.0%
Tolerance
3
Runs
3
Passed

Calibrated from GEE execution on 2026-03-29 (actual: 0.199). Will be validated against QGIS and folia outputs. Variation expected from cloud masking aggressiveness, fill-value handling, and surface reflectance scaling factors across platforms. Landsat C2 L2 scale factor is 0.0000275 with offset -0.2.

Run Results

Workflow Model Backend Status Answer Error Cost Latency
exec gold gee PASS 0.19911882722697605 0.4% $0.001542 27.8s
exec gold folia-rust PASS 0.17075136303901672 14.6% --- 17ms
exec gold qgis PASS 0.1708 14.6% --- 200ms

Requirements

Data Sources
  • Landsat 8 Collection 2 Level 2 Surface Reflectance
Operations
cloud_mask_landsat imagery_composite temporal_reduce

LLM Prompt

The prompt given to LLMs in single-shot workflow benchmarks. 

Problem: Create an annual median composite from Landsat 8 Collection 2
Level 2 imagery for 2022 over Yellowstone National Park. Apply cloud
masking and report the median B5 (NIR) reflectance value.

Study area: -111.0, 44.3, -110.5, 44.8 (Yellowstone NP).
Data: Landsat 8 C2 L2 (LANDSAT/LC08/C02/T1_L2).
Expected answer: approximately 0.20 median NIR reflectance.