Conversion of Horizontal to Plane-of-Array Irradiance: (When) Is the "Liu and Jordan" Albedo Assumption Justified?

Plane-of-array (POA) solar irradiance values are frequently needed for the estimation
of solar panel production. Because POA measurements are uncommon, this often involves
transformation of measured total (global) horizontal solar irradiance values to POA
using any of several models to decompose the measurement into direct and diffuse
terms and then transpose these terms to match the alignment of the receiving plane.
Surface albedo is an important factor in this conversion, since it controls the
fraction of the light striking the surface that scatters back onto the tilted plane.
Unfortunately, surface albedo is also frequently unknown.

Since the 1960s, an assumed surface albedo of 0.2 has frequently been applied in
such applications. This value comes from a simple expression meant to represent
the properties at the Blue Hill, MA, observatory, given in work by Benjamin Liu and
Richard Jordan. However, the reflective characteristics of soils and vegetation
vary widely, with values from less than 0.1 to more than 0.35 appearing in the
literature. Man-made surfaces cover a larger range, reaching values above 0.5,
while the albedo of fresh snow of sufficient depth can exceed 0.9. It is therefore
necessary to examine the variability of albedo with location and over time
at individual locations.

We will illustrate the range of albedos found under a variety of outdoor conditions
using measurements from stations in a range of climate regions. The impact of
variations in albedo on computed POA irradiance will then be assessed using typical
plane-of-array conversion algorithms. These results will account for changes as a
function of panel orientation and time of year and indicate the level of error that
results from assuming an albedo of 0.2.