Geoscience Reference
In-Depth Information
In a later development, Calder (1996) extended the stochastic model to account for sec-
ondary drops, that is drops falling from upper parts of the canopy that strike lower elemental
areas, and for a kinetic energy dependence effect that is a function of drop size and rainfall in-
tensity. This version of the model has been tested by Calder
et al.
(1996) and Hall
et al.
(1996).
This model requires seven parameters to be specified. This is too many to be determined from
measurements of throughfall rates alone and calibration requires some detailed measurements
during wetting experiments, including drop sizes (Calder
et al.
, 1996).
Box 3.3 Estimating Snowmelt by the Degree-Day Method
As noted in the main text, there are many problems and complexities in modelling the accu-
mulation and melting of snowpacks on a complex topography. While process-based energy
budget and pack evolution models are available (e.g. Anderson, 1968; Bl oschl
et al.
, 1991;
Morris, 1991; Marks and Dozier, 1992), it has proven difficult to demonstrate that they can
produce generally more accurate operational predictions than simpler empirical models, at
least without assuming that parameter values, such as albedo of the pack, are time variable
(e.g. Braun and Lang, 1986). Of the empirical models, the most widely used is the degree-day
method. An early version of the degree-day method was presented by Imbeaux (1892), includ-
ing the use of different air temperatures in different elevation bands in predicting snowmelt
floods in the Durance catchment in France. There are now many modern variants of the ap-
proach (e.g. Bergstrom, 1975; Martinec and Rango, 1981; Hottelet
et al.
, 1993; Moore
et al.
,
1999; Hock, 2003; Zhang
et al.
, 2006). An example of the accuracy of snowmelt discharge
predictions in a large catchment for two years using a degree-day snowmelt model with a daily
time step is shown in Figure B3.3.1.
Figure B3.3.1
Discharge predictions for the Rio Grande basin at Del Norte, Colorado (3419 km
2
)
using the Snowmelt Runoff model (SRM) based on the degree-day method (after Rango, 1995, with
kind permission of Water Resource Publications).
