Atmospheric Observation, Remote Sensing and Numerical Weather Prediction
Atmospheric observation, remote sensing, and numerical weather prediction covers radiosondes, radar, satellite channels, data assimilation, and model behavior. The exam tests instrument interpretation and the limits of NWP, not just definitions.
Core rules
- Radiosondes measure T, RH, and wind with height; balloon drift and sensor lag are known error sources. GNSS-RO provides refractivity and temperature information with high vertical resolution.
- Radar reflectivity factor is Z = Σ D⁶ (mm⁶ m⁻³), reported as dBZ = 10 log10 Z. Doppler velocity gives radial motion; range folding and velocity folding are common artifacts.
- Satellite channels sense different layers: visible reflects sunlight, infrared senses cloud-top temperature, and water vapor senses upper-tropospheric moisture.
- NWP data assimilation (3D-Var, 4D-Var, EnKF) blends observations with a background forecast; parameterizations represent unresolved processes. Ensembles sample initial-condition and model uncertainty.
Worked example
For Marshall-Palmer Z = 200 R^1.6 and a reported 40 dBZ, first convert: Z = 10^(40/10) = 10,000 mm⁶ m⁻³. Then R = (Z/200)^(1/1.6) = 50^0.625. Since ln 50 = 3.912 and 3.912 × 0.625 = 2.445, R ≈ 11.6 mm h⁻¹. Note that a 10-dBZ increase is a tenfold increase in Z, not a tenfold increase in rain rate.
Traps
Treating dBZ as linear; confusing reflectivity with rain rate; ignoring the bright band; reading radial velocity as total wind; assuming more data assimilation always fixes model error; and using a single deterministic run as if it were a probability forecast.