Evaluation of a new 3-D radiative transfer model to simulate snow albedo over a macroscopic rough surface
Abstract
Snow spectral albedo is one of the most important parameters of the energy budget at the surface, and a good modeling and an understanding of this variable is crucial for climate change models. Albedo varies according to snow properties (with the Specific Surface Area of snow (SSA), impurities, etc.) and to the spectral and angular signatures of the solar incident radiation. Radiative transfer models were developed to simulate snow albedo according to snowpack evolution and are often used to retrieve SSA values by inversion methods using albedo observations. Nevertheless, these models assume plane-parallel layers with smooth and flat surfaces. In reality, the macroscopic surface roughness leads to a reduction in snow albedo due to multiple reflections and a 'trapping' effect of photons. Several studies developed models to reproduce these effects on the spectral reflectance of snow but most of them focus on the bi-directional reflectance distribution function and not on snow albedo. However, the roughness effects can introduce large uncertainties in snow albedo simulations if they are not well represented.
We propose a new Monte Carlo photon transport algorithm (MC model) to simulate snow surface albedo over a macroscopic rough surface. The MC model considers both the geometric effects owing to roughness with a 3-D mesh of the studied area and the optical snow properties through analytical equations. Measurements were acquired in the French Alps during the 2017/2018 snow season for different illumination conditions and for a snow surface where roughness was artificially created. The use of a 'controlled' environment allowed us to quantify the different contributions on snow albedo measurements (illumination conditions, SSA, roughness). Measurements over a rough surface show an average decrease in albedo of 5 % compared to a smooth surface. The difference can reach 7 % when the roughness axis is perpendicular to the sun. Depending on the number and the shape of roughness features, the values of optimised SSA from albedo measurements can be underestimated by 50 % using models with flat surfaces. The in situ data were subsequently compared to albedo simulations showing a good agreement with the MC model. By considering surface roughness, uncertainties of optimised SSA are clearly reduced.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.C41B..06L
- Keywords:
-
- 0736 Snow;
- CRYOSPHEREDE: 0740 Snowmelt;
- CRYOSPHEREDE: 0798 Modeling;
- CRYOSPHERE