plays a critical role in flash flood prediction. These observational challenges posed to NEXRAD in complex terrain and coastal regions—for example, in Southern California’s Ventura and Los Angeles Counties—are discussed in more detail in this chapter.

Propagation of radar waves in the atmosphere is a well-understood process (Bean and Dutton, 1966). If the atmospheric properties such as temperature, humidity, and pressure are known, the propagation path of the radar beam is easy to characterize. Design studies of radar networks usually assume a standard atmosphere (Doviak and Zrnic, 1985). Assuming that the beam propagates under standard atmospheric conditions allows determination of the potential interactions of the beam with the surrounding terrain. To achieve this, two other factors must be considered. First, the radar beam spreads as it propagates away from the radar so that the NEXRAD beam is more than 1 km wide at a distance of about 60 km from the radar site. Second, the power distribution within the radar beam is not uniform; most of the power is concentrated near the main axis of the beam, and the intensity falls off approximately following a Gaussian pattern. Thus, to properly determine the extent to which terrain affects radar observations, detailed information about the terrain, location, and hardware parameters of the radar and its scanning pattern (i.e., antenna elevation angle) must be known in addition to atmospheric conditions.

Complete blockage of the radar beam by terrain features occurring at a given distance and direction prevents detection of weather echoes past the location of the blockage. The blocking element (i.e., hills, mountains, nearby buildings) leads to a strong and permanent echo that is visible on the radar display. If the radar beam is only partially blocked, it may still be useful for observing weather echoes if the amount of power remaining in the beam is large enough for the radar return to be above the signal detection level. However, precipitation observations collected by partially blocked radar beams may lead to biased results if left uncorrected. Corrections for partial beam blockage can be part of radar data processing algorithms, but correcting the data often is difficult because the degree to which the beam is obstructed rarely is known accurately (Seo et al., 2000; Kucera et al., 2004). In addition, the corrections may rely on assumptions that are difficult to verify because complex terrain modifies observed echo patterns (Warner et al., 2000; Ralph et al., 2003).