Examination of sub-pixel accumulation variability in central Nebraska

To try to bound the error on Z-R relationships, the rainfall inversion is often checked against rain gauges to "ground truth" the process. Recent studies in the microphysics of precipitation, however, have suggested that rainfall has coherent structure on brief (less than 5 minutes) and small (less than 1 km) time and spatial scales. Although it is often noted that the precipitation measured aloft by a radar is not the same as the rain observed by gauges on the ground, there are often the implicit assumptions that (1) the bulk rainfall information inferred for each pixel in a scan is at least approximately true for the duration of the scan and spatial extent of that pixel or (2) that it at least "averages out" over the duration of the rain event.

We have examined these hypotheses by constructing two rain gauge arrays, each array having all of its gauges located within a single radar pixel. The rainfall accumulation statistics of these arrays are then compared to the Z-R inferred rain-rate using a representative Z-R relationships for the entire pixel and neighboring pixels. Our results cast doubt on the validity of the universality of these common assumptions.