WMOD00472

• Title: Estimation of Cloud Liquid Water in Winter Storms on the Mogollon Rim
• Date: 5/1/1993
• Weather Modification - Doc Type: Report

Developing CLW estimation capability is a formidable task because of substantial natural
variability of storms and incomplete knowledge and understanding of the physics involved.
Thorough and general solutions with widespread application will require years of study and
analysis. However, the possibility that some important features of Mogollon Rim winter storms
may have sufficient regularity, such that useful relationships providing forecasts of CLW
and/or CLW statistics (based on those features) could be developed from historical data, led to
the conduct of the research reported here. The rationale is that forecast CLW estimates and
related measured precipitation could then be employed to develop other relationships that
estimate cloud PE. Because the object of cloud seeding is to improve PE, relationships with
apparent predictive skill could be applied to a lengthy historical data set to estimate the
feasibility of cloud seeding on dry, normal, and wet winters.

In the initial task (task 1) of the study, the SRP requested that efforts be directed at the
development of relationships that estimate CLW in a historical data set, or at a minimum,
CLW statistics (e.g., a specific category of storms or days produce a given interval amount of
CLW) from such data. The goal was not the real-time forecasting of CLW for cloud seeding
decisions. Background information would consist of available historical data from national
archives, the 1987 experiments, and simulations from the precipitation model adapted to the
Mogollon Rim. Provided the initial research task met with sufficient success, a second task
(task 2) is to be considered for future accomplishment, aimed at pursuing the development of
relationships that associate PE with CLW.

The following section contains discussion of a statistical approach to determining relationships
that estimate CLW. The subsequent section covers the adaptation and application of a
one-dimensional model to estimate CLW from soundings. Finally, the last section discusses the
implications ofthe research.

STATISTICAL APPROACH

Methodology

The initial approach considered for developing relationships that estimate CLWas measured
by a radiometer was largely statistical in nature. Generally, the approach consisted of
developing scatterplots of the available CLW measurements versus individual sounding
variables, and inspecting the results for apparent linear or other recognizable relationships.
With linear or transformable to linear relationships apparent in some of the sounding
variables, those appearing best related to CLW were used in the application MLR (multiple
linear regression). The sounding variables were the independent variables and CLW posed as
the dependent variable in the MLR. Desirable results were that sounding variables, expressed
in a simplified or some transformed mode, explain noticeable amounts of the variance ofCLW.

Ground truth measurements of CL W were obtained from analysis of data collected during the
1987 experiments by the vertically-pointing radiometer located at HJ (Happy Jack) (fig. 1).
Local sounding data were available from rawinsondes launched at CVR (Camp Verde) at time
intervals of as few as 3 hours.

The resulting MLR equation would estimate HJ CLW measurements. Application of the
equation to available sounding data for other winters that included wet, normal, and dry

2