Measured and satellite-derived albedo data for estimating bifacial photovoltaic system performance

Abstract

The albedo of the ground surface is an important factor in the cost-effectiveness of a bifacial photovoltaic (PV) system. To improve the availability of reliable albedo data, datasets of ground albedo and associated meteorological data were developed by using existing measurement network data and data measured by the PV industry. The measured datasets include time-series data as well as summary information of tabular monthly and yearly data and plots of monthly and hourly albedo values. Satellite-derived values of albedo are available from the National Solar Radiation Data Base (NSRDB). The NSRDB albedos were compared to the measured albedos for Surface Radiation budget (SURFRAD) network locations for the period 2001–2017, and the mean bias difference results were from −0.044 to +0.056. These differences are greater than the albedo measurement uncertainty of ±0.02; consequently, the NSRDB albedos should be used with caution for estimating the performance of bifacial PV systems. Differences between SURFRAD and NSRDB albedos are attributed to the NSRDB method for determining albedo and to the ground surfaces within the NSRDB 4 km spatial resolution pixel consisting of a mixture of surface types rather than just the single surface types viewed by the albedometers at the SURFRAD stations.

Introduction

The glass/glass or glass/transparent backsheet construction of a bifacial photovoltaic (PV) module permits its cells to convert to electricity the solar radiation received by both the front and back sides of the PV module. Compared to monofacial PV systems, the increased energy may be modest, but because the incremental cost may be small, there is considerable interest in the deployment of large-scale bifacial PV systems. For dusty locations, another benefit is a smaller energy loss from soiling. For example, in Santiago, Chile, the soiling rate was 8.8 times smaller for the backside of a PV module compared to the front side (Luque et al., 2018), and in Mumbai, India, the soiling loss for a vertical bifacial PV module was essentially nonexistent (Bhaduri and Kottantharayil, 2019).

The widespread use of bifacial PV modules requires standards development for their indoor characterization and performance models to predict their outdoor performance (Liang et al., 2019). Compared to modeling the performance of a monofacial PV system, bifacial PV system performance modeling has the added complexity of determining the solar radiation received by the backside of the PV module. The main source of the solar radiation received by the backside of the PV module is that reflected from the ground surface. There is also a small amount received from the sky, and for equator-facing tilted PV modules, there may be solar radiation received directly from the sun’s disk during early morning or late evening hours during the summer.

The PV module backside irradiance is modeled using either ray tracing or view-factor principles. Because the amount of radiation reflected from the shadowed ground is much less than for the sunlit ground, an important factor is to correctly determine the extent of the ground shadowed by the PV array. Several public and commercial models are available for modeling the performance of bifacial PV systems. These models provide reasonable results (Chudinzow et al., 2019, Pelaez et al., 2019, Nussbaumer et al., 2020).

A needed input for bifacial PV system performance models is the albedo of the ground surface, whose value has been shown to be an important factor for determining if a bifacial PV system is more cost-effective than a monofacial PV system (Rodriguez-Gallegos et al., 2018). Albedo is the fraction of the incident sunlight that the surface reflects. It is not strictly a fixed property of a surface because the amount reflected is subject to angular and spectral effects as the sun position, solar spectrum, and the diffuse and beam proportions change over the course of a day, and with season and latitude. Except for ice, snow, and water, most natural surfaces exhibit low albedos in the visible wavelengths and an abrupt increase at about 700 nm (Iqbal, 1983). This is most evident for green vegetation that uses the photosynthetically active radiation from 400 nm to 700 nm for plant growth.

The albedo for bare soils depends on the soil moisture and surface roughness. For example, the albedo of a loam soil in Phoenix, Arizona, changed from 0.14 when wet to 0.30 when dry, with a linear relationship based on moisture content (Idso et al., 1975). The effect of surface roughness was studied by Matthias et al. (2000) using four different farm tillage conditions (rough-plow, disk, disk-disk, and seedbed, with rough-plow being the roughest condition) and for two soil types in Tucson, Arizona. Compared to a smooth soil condition, the four tillage conditions reduced the albedo by 27%, 18%, 10%, and 8%, respectively. The albedos are less for the rougher surfaces primarily because of self-shading. (In the southwest U.S., the tillage practice is sometimes selected to alter the soil albedo to optimize soil temperatures for agricultural purposes.)

In the United States, measurement networks that measure albedo include the Surface Radiation budget (SURFRAD) network and the AmeriFlux network. The SURFRAD network consists of seven stations and is operated by the National Oceanic and Atmospheric Administration (NOAA) to provide continuous and high-quality surface radiation budget measurements to support climate research, weather forecasting, satellite, and educational communities (Augustine, 2000). The AmeriFlux network data are contributed by individual scientists that operate measurement stations in North, Central, and South America for the purpose of measuring ecosystem carbon dioxide, water, and energy fluxes. Most of the sites are in the United States, followed by Canada. The AmeriFlux network is managed by the Lawrence Berkeley National Laboratory with funding from the U.S. Department of Energy’s Office of Science (AmeriFlux, 2020). The AmeriFlux network is intended to represent major climate and ecological biomes including tundra, grasslands, savanna, crops, and forests.

Numerous satellite-derived albedo databases exist with varying spatial and temporal resolution. Albedo is an essential element for determining the earth’s energy balance and climate change, and satellite-derived albedos provide data with the global coverage required. Gueymard et al. (2019) identifies 21 databases with satellite-derived albedos. An important source of albedo data is the Moderate Resolution Imaging Spectroradiometer (MODIS) data measured with sensors onboard Terra and Aqua satellites, beginning in 2001. Albedo products from these data are derived from multiangle measurements of surface reflectance over 16-day periods when skies are clear (Schaaf et al., 2002). The anisotropy of the surface reflectance is quantified by determining model parameters for the Bidirectional Reflectance Distribution Function (BRDF). The BRDF describes mathematically the changes in reflectance observed when an illuminated surface is viewed from different angles. Various studies demonstrate the effectiveness of this approach in estimating the ground surface albedo (Knobelspiesse et al., 2008, Wang et al., 2015, Liu et al., 2017).

For monofacial PV systems, the ground-reflected radiation typically comprises only 1% to 2% of the total radiation received by the PV module. Consequently, a rudimentary understanding of the ground albedo is adequate for predicting their performance. However, for bifacial PV modules where their benefit is determined by the additional radiation reflected by the ground to their backside, a better understanding of albedo values and characteristics is needed by both the PV and financial communities to better estimate performance and to reduce risk.

To meet this need, the National Renewable Energy Laboratory (NREL) initiated an effort to provide albedo data for different locations and ground surfaces for both the simulation of bifacial PV systems and for summary information, including variations in albedo with respect to time of day, season, and year. The resulting datasets include data from the SURFRAD and AmeriFlux networks, and data contributed by the PV industry.

The objective of this work is to provide information on the measured albedo datasets and to compare their values with satellite-derived albedos to better understand the applicability of both measured and satellite-derived albedos for modeling the performance of bifacial PV systems. In Section 2, the measured albedo datasets, their characteristics, and the available data and summary statistics are described; in Section 3, the method for determining the satellite-derived albedos for the National Solar Radiation Data Base (NSRDB) is summarized; in Section 4, the satellite-derived albedos from the NSRDB are compared with the measured albedos from the SURFRAD network for the same locations; and in Section 5, the findings are summarized and future work is identified.