Review article: Parameterizations of snow-related physical processes in land surface models

Snow on land surface plays a vital role in the interaction between land and atmosphere in the state-of-the-art land surface models (LSMs) and the real world. Since the snow cover affects the snow albedo and the ground and soil heat fluxes, it is crucial to detect snow cover changes accurately. It is challenging to acquire observation data for snow cover, snow albedo, and snow depth; thus, an excellent alternative is to use the simulation data produced by the LSMs that calculate the snow-related physical processes. The LSMs show significant differences in the complexities of the snow parameterizations in terms of variables and processes considered. Thus, the synthetic intercomparisons of the snow physics in the LSMs will help the improvement of each LSM. This study revealed and discussed the differences in the parameterizations among LSMs related to snow cover fraction, snow albedo, and snow density. We selected the most popular and well-documented LSMs embedded in the Earth System Model or operational forecasting systems. We examined single layer schemes, including the Unified Noah Land Surface Model (Noah LSM), the Hydrology Tiled ECMWF Scheme of Surface Exchanges over Land (HTESSEL), the Biosphere-Atmosphere Transfer Scheme (BATS), the Canadian Land Surface Scheme (CLASS), and multilayer schemes of intermediate complexity including the Community Noah Land Surface Model with Multi-Parameterization Options (Noah-MP), the Community Land Model version 5 (CLM 5), the Joint UK Land Environment Simulator (JULES), and the Interaction Soil-Biosphere-Atmosphere (ISBA). First, we identified that BATS, Noah-MP, JULES, and ISBA reflect the snow depth and roughness length to parameterize snow cover fraction, and CLM 5 accounts for the standard deviation of the elevation value for the snow cover decay function. Second, CLM 5 and BATS are relatively complex, so that they explicitly take into account the solar zenith angle, black carbon, mineral dust, organic carbon, and ice grain size for the determinations of snow albedo. Besides, JULES and ISBA are also complicated model which concerns ice grain size, solar zenith angle, new snow depth, fresh snowfall rate, and surface temperature for the albedo scheme. Third, HTESSEL, CLM 5, and ISBA considered the effects of both wind and temperature in the determinations of the new snow density. Especially, ISBA and JULES considered internal snow characteristics such as snow viscosity, snow temperature, and vertical stress for parameterizing new snow density. The future outlook discussed geomorphic and vegetation-related variables for the further improvement of the LSMs. Previous studies clearly show that spatio-temporal variation of snow is due to the influence of altitude, slope, and vegetation condition. Therefore, we recommended applying geomorphic and vegetation factors such as elevation, slope, time-varying roughness length, vegetation indexes, or optimized parameters according to the land surface type to parameterize snow-related physical processes.

Optimization of snow density parameter of Noah Land Surface Model using micro-genetic algorithm for estimating snow depth

<p>The Noah Land Surface Model (Noah LSM) estimates snow depth using snow water equivalent and snow density. The snow density is determined by snow compaction, snowmelt water storing, and density of fresh snowfall. The Noah LSM usually underestimates snow depth compared to the ground observations in Korea, which occurs from the beginning of snowfall. We performed an optimal estimation of parameters related to the density of fresh snowfall, using micro-genetic algorithm (μ-GA) that uses the evolution process concept through natural selection and mutation mechanism. Ground observations from 36 sites of the Korea Meteorological Administration, for the recent 10 years (May 2009 – April 2019), are used for offline forcing of the Noah LSM and evaluating the fitness function in μ-GA. Optimized parameters reduced the density of fresh snowfall, and improved the simulated snow depth. The root-mean-square error of snow depth decreased from 8.1 cm to 7.1 cm.</p>

Impact of GVF Derivation Methods on Noah Land Surface Model Simulations and WRF Model Forecasts

Green vegetation fraction (GVF) plays a crucial role in the atmosphere–land water and energy exchanges. It is one of the essential parameters in the Noah land surface model (LSM) that serves as the land component of a number of operational numerical weather prediction models at the National Centers for Environmental Prediction (NCEP) of NOAA. The satellite GVF products used in NCEP models are derived from a simple linear conversion of either the normalized difference vegetation index (NDVI) from the Advanced Very High Resolution Radiometer (AVHRR) currently or the enhanced vegetation index (EVI) from the Visible Infrared Imaging Radiometer Suite (VIIRS) planned for the near future. Since the NDVI or EVI is a simple spectral index of vegetation cover, GVFs derived from them may lack the biophysical meaning required in the Noah LSM. Moreover, the NDVI- or EVI-based GVF data products may be systematically biased over densely vegetated regions resulting from the saturation issue associated with spectral vegetation indices. On the other hand, the GVF is physically related to the leaf area index (LAI), and thus it could be beneficial to derive GVF from LAI data products. In this paper, the EVI-based and the LAI-based GVF derivation methods are mathematically analyzed and are found to be significantly different from each other. Impacts of GVF differences on the Noah LSM simulations and on weather forecasts of the Weather Research and Forecasting (WRF) Model are further assessed. Results indicate that LAI-based GVF outperforms the EVI-based one when used in both the offline Noah LSM and WRF Model.

Impact of Using Near Real-Time Green Vegetation Fraction in Noah Land Surface Model of NOAA NCEP on Numerical Weather Predictions

Green vegetation fraction (GVF) is one of the input parameters of the Noah land surface model (LSM) that is the land component of a number of operational numerical weather prediction (NWP) models at the National Centers for Environmental Prediction (NCEP) of NOAA. The Noah LSM in current NCEP operational NWP models has been using static multiyear averages of monthly GVF derived from satellite observations of NOAA Advanced Very High Resolution Radiometer (AVHRR) normalized difference vegetation index. The multiyear averages of GVF are evidently not the representative of actual conditions of the land surface vegetation cover. This study used a near-real-time (NRT) GVF data set generated from the 8-day composite of the leaf area index product from the Moderate Resolution Imaging Spectroradiometer (MODIS) to assess the impact of NRT GVF on off-line Noah LSM simulations and NWP forecast model. Simulations of the off-line Noah LSM in the Land Information System (LIS) and weather forecasts of the NASA-Unified Weather and Research Forecasting (NUWRF) were obtained using either the static multiyear average AVHRR GVF data set or the NRT MODIS GVF while meteorological forcing data and other settings were kept the same. The off-line simulations and WRF forecasts were then compared against in situ measurements or reanalysis products to assess the impact of using NRT GVF. Improvements of both soil moisture simulations as well as forecasts of 2-meter air temperature and humidity and precipitation from NUWRF were observed using the NRT GVF data products. The RMSE in SM estimates from the off-line Noah model is reduced by around 1.0% (1.41%) during the green-up phase and by 1.48% (2.24%) over the senescence phase for the surface (root zone) SM simulations. Around 82.3% validation sites (out of 1178 sites) showed positive impact on coupled WRF model with the insertion of NRT GVF.

Evaluation of Snow Water Equivalent in NARCCAP Simulations, Including Measures of Observational Uncertainty

This study evaluates snow water equivalent (SWE) over North America in the reanalysis-driven NARCCAP regional climate model (RCM) experiments. Examination of SWE in these runs allows for the identification of bias due to RCM configuration, separate from inherited GCM bias. SWE from the models is compared to SWE from a new ensemble observational product to evaluate the RCMs’ ability to capture the magnitude, spatial distribution, duration, and timing of the snow season. This new dataset includes data from 14 different sources in five different types. Consideration of the associated uncertainty in observed SWE strongly influences the appearance of bias in RCM-generated SWE. Of the six NARCCAP RCMs, the version of MM5 run by Iowa State University (MM5I) is found to best represent SWE despite its use of the Noah land surface model. CRCM overestimates SWE because of cold temperature biases and surface temperature parameterization options, while RegCM3 (RCM3) does so because of excessive precipitation. HadRM3 (HRM3) underestimates SWE because of warm temperature biases, while in the version of WRF using the Grell scheme (WRFG) and ECPC-RSM (ECP2), the misrepresentation of snow in the Noah land surface model plays the dominant role in SWE bias, particularly in ECP2 where sublimation is too high.