A random molecular process where there is net movement of variables such as chemicals, temperature, moisture, or momentum, from regions of high concentration toward lower concentration. A classic example is heat flowing from the hot end toward the cold end of a metal rod. Gradient refers to the change of mean variable across a distance, such as the temperature change across the metal rod. Down-gradient refers to movement from high to low concentration, such as from hot to cold (as opposed to counter-gradient, meaning from cold to hot). Many turbulence theories such as K-theory and mixing-length theory assume that turbulent transport is caused mostly by small eddies, resulting in a local transport (down-gradient turbulent diffusion). Such small eddy local turbulent transport is not a law, but is a turbulence closure assumption that works well for those atmospheric situations where small eddies dominate (such as in the atmospheric surface layer), but that usually fails for situations where large eddies dominate (such as in the atmospheric mixed layer where large convective thermals cause nonlocal turbulent transport). Compare counter-gradient flux.