Implementation and comparison of a suite of heat stress metrics within the Community Land Model version 4.5

Abstract

Abstract. We implement and analyze 13 different metrics (4 moist thermodynamic quantities and 9 heat stress metrics) in the Community Land Model (CLM4.5), the land surface component of the Community Earth System Model (CESM). We call these routines the HumanIndexMod. These heat stress metrics embody three philosophical approaches: comfort, physiology, and empirically based algorithms. The metrics are directly connected to CLM4.5 BareGroundFuxesMod, CanopyFluxesMod, SlakeFluxesMod, and UrbanMod modules in order to differentiate between the distinct regimes even within one gridcell. This allows CLM4.5 to calculate the instantaneous heat stress at every model time step, for every land surface type, capturing all aspects of non-linearity in moisture-temperature covariance. Secondary modules for initialization and archiving are modified to generate the metrics as standard output. All of the metrics implemented depend on the covariance of near surface atmospheric variables: temperature, pressure, and humidity. Accurate wet bulb temperatures are critical for quantifying heat stress (used by 5 of the 9 heat stress metrics). Unfortunately, moist thermodynamic calculations for calculating accurate wet bulb temperatures are not in CLM4.5. To remedy this, we incorporated comprehensive water vapor calculations into CLM4.5. The three advantages of adding these metrics to CLM4.5 are (1) improved thermodynamic calculations within climate models, (2) quantifying human heat stress, and (3) that these metrics may be applied to other animals as well as industrial applications. Additionally, an offline version of the HumanIndexMod is available for applications with weather and climate datasets. Examples of such applications are the high temporal resolution CMIP5 archived data, weather and research forecasting models, CLM4.5 flux tower simulations (or other land surface model validation studies), and local weather station data analysis. To demonstrate the capabilities of the HumanIndexMod, we analyze the top 1% of heat stress events from 1901–2010 at a 4 × daily resolution from a global CLM4.5 simulation. We cross compare these events to the input moisture and temperature conditions, and with each metric. Our results show that heat stress may be divided into two regimes: arid and non-arid. The highest heat stress values are in areas with strong convection (±30° latitude). Equatorial regions have low variability in heat stress values (±20° latitude). Arid regions have large variability in extreme heat stress as compared to the low latitudes.