Future extreme hourly wet bulb temperatures using downscaled climate model projections of temperature and relative humidity

Abstract

Extreme wet bulb temperatures are important for a number of applications including the proper and efficient design of building cooling systems. Since wet bulb temperature is not directly available from climate model output and design specifications require information at hourly resolution, whereas twice-daily resolution is more typical of climate models, the ability of climate models to replicate the observed climatology is evaluated at a set of US stations. Observed wet bulb extremes can be replicated by applying a spline fit to the twice-daily humidity and temperature observations that simulate the data available from climate models and then minimizing the residual of the equation specifying the change in enthalpy of moist air. On average, these ersatz values are 1 °C colder than the observed values. Climate model simulations for the period 1950–2005 also generally agree with the ersatz observations. At most locations, the model bias is negative (model values colder than the simulated observations) and on average near 1 °C. The largest positive biases occur at the most arid stations and the largest negative biases are found at the coldest locations. Model projections for the mid-twenty-first century indicate that the most extreme wet bulb temperatures will increase by between 1 and 2.3 °C, with the largest increases at the most northern locations. Future warming and wetting appear to result in a translation of the entire wet bulb cumulative distribution function, leading to similar increases regardless of wet bulb temperature. The increase is fairly consistent among different climate models and at each station.

Appendices

Appendix 1. List of climate model acronyms

Acronym :

Modeling group

ACCESS1.0:

CSIRO, Australia

ACCESS1.3:

CSIRO, Australia

BCC−CSM1.1:

Beijing Climate Center, China

BCC−CSM1.1(m):

Beijing Climate Center, China

CCSM4:

National Center for Atmospheric Research, USA

CNRM−CM5:

National Center for Meteorological Research, France

CanESM2:

Canadian Center for Climate Modeling and Analysis, Canada

FGOALS−g2:

LASG, China

GFDL−CM3:

Geophysical Fluid Dynamics Laboratory, USA

GFDL−ESM2G:

Geophysical Fluid Dynamics Laboratory, USA

GFDL−ESM2M:

Geophysical Fluid Dynamics Laboratory, USA

GISS−E2−H:

NASA Goddard Institute for Space Science, USA

GISS−E2−R:

NASA Goddard Institute for Space Science, USA

HADGEM2−AO:

Met Office Hadley Centre, UK

HADGEM2−CC:

Met Office Hadley Centre, UK

HADGEM2−ES:

Met Office Hadley Centre, UK

inmcm4:

Institute for Numerical Mathematics, Russia

IPSL−CM5A:

Pierre Simon Laplace Institute, France

IPSL−CM5A−MR:

Pierre Simon Laplace Institute, France

MIROC−ESM:

JAMSTEC/AORI/NIES, Japan

MIROC−ESM−CHEM:

JAMSTEC/AORI/NIES, Japan

MIROC5:

JAMSTEC/AORI/NIES, Japan

MRI−CGCM3:

Meteorological Research Institute, Japan

NorESM1−M:

Norwegian Climate Center, Norway

Appendix 2. Weather station abbreviations

KABQ:

Albuquerque, NM

KBNA:

Nashville, TN

KBOS:

Boston, MA

KBTV:

Burlington, VT

KBUF:

Buffalo, NY

KCAR:

Caribou, ME

KDCA:

Washington, DC

KDFW:

Dallas, TX

KFAR:

Fargo, ND

KGTF:

Great Falls, MT

KLNK:

Lincoln, NE

KMIA:

Miami, FL

KMSY:

New Orleans, LA

KORD:

Chicago, IL

KRAP:

Rapid City, SD

KSAN:

San Diego, CA

KSFO:

San Francisco, CA

KSEA:

Seattle, WA

KTUS:

Tucson, AZ