Models that predict meltwater runoff at a daily timescale are important in
water resource management, flood hazard assessment and climate-change impac
t studies. This article identifies four basic components of such models: me
teorological extrapolation, snowmelt estimation at a point, snow-cover depl
etion and runoff routing. Alternative ways of handling these are discussed,
with emphasis on the contrasting treatments in two widely used models: HBV
and SRM. Many of the issues in meltwater modelling reflect wider debates i
n hydrological and environmental modelling, including problems of complexit
y vs. simplicity, the appropriate level of spatial disaggregation, paramete
r identification and calibration, and internal validation. In reviewing cur
rent trends emphasis is placed on the potential and Limitations of fully di
stributed models, problems in using energy-balance rather than temperature-
index melt models at basin scale, ways to deal with spatial variability in
snow cover, and the value and limitations of earth observation data.