Is the past a guide to the future? Evaluating the assumption of climate stationarity in statistical downscaling

Katharine Hayhoe; Keith Dixon; Anne Stoner; John Lanzante; Aparna Radhakrishnan

Is the past a guide to the future? Evaluating the assumption of climate stationarity in statistical downscaling

Katharine Hayhoe, Keith Dixon, Anne Stoner, John Lanzante, Aparna Radhakrishnan

Political Science

Research output: Contribution to conference › Paper

Abstract

Climate change requires a global perspective to understand the past and explore the future. The impacts of climate change, however, are experienced mainly at the local to regional level. Downscaling techniques are commonly used to bridge the gap between the spatial scales at which climate is modeled vs. the scales at which impact assessments require climate projections. One of the most common assumptions made by downscaling is that of stationarity: that current-day relationships between climate variables, relationships that cannot be directly represented by fundamental physical equations but rather must be parameterized or statistically modeled, hold true under very different future conditions. As future observations are not yet available, the validity of this assumption is difficult to test. Here, by treating 25km output from the GFDL-HiRAM-C360 model as “observations” for both past and future periods, we quantify the ability of three different statistical downscaling methods (season

Original language	English
State	Published - Dec 6 2012

Fingerprint Dive into the research topics of 'Is the past a guide to the future? Evaluating the assumption of climate stationarity in statistical downscaling'. Together they form a unique fingerprint.

Cite this

@conference{53450415fcd546a69090c27c00d32428,

title = "Is the past a guide to the future? Evaluating the assumption of climate stationarity in statistical downscaling",

abstract = "Climate change requires a global perspective to understand the past and explore the future. The impacts of climate change, however, are experienced mainly at the local to regional level. Downscaling techniques are commonly used to bridge the gap between the spatial scales at which climate is modeled vs. the scales at which impact assessments require climate projections. One of the most common assumptions made by downscaling is that of stationarity: that current-day relationships between climate variables, relationships that cannot be directly represented by fundamental physical equations but rather must be parameterized or statistically modeled, hold true under very different future conditions. As future observations are not yet available, the validity of this assumption is difficult to test. Here, by treating 25km output from the GFDL-HiRAM-C360 model as “observations” for both past and future periods, we quantify the ability of three different statistical downscaling methods (season",

author = "Katharine Hayhoe and Keith Dixon and Anne Stoner and John Lanzante and Aparna Radhakrishnan",

year = "2012",

month = dec,

day = "6",

language = "English",

}

TY - CONF

T1 - Is the past a guide to the future? Evaluating the assumption of climate stationarity in statistical downscaling

AU - Hayhoe, Katharine

AU - Dixon, Keith

AU - Stoner, Anne

AU - Lanzante, John

AU - Radhakrishnan, Aparna

PY - 2012/12/6

Y1 - 2012/12/6

N2 - Climate change requires a global perspective to understand the past and explore the future. The impacts of climate change, however, are experienced mainly at the local to regional level. Downscaling techniques are commonly used to bridge the gap between the spatial scales at which climate is modeled vs. the scales at which impact assessments require climate projections. One of the most common assumptions made by downscaling is that of stationarity: that current-day relationships between climate variables, relationships that cannot be directly represented by fundamental physical equations but rather must be parameterized or statistically modeled, hold true under very different future conditions. As future observations are not yet available, the validity of this assumption is difficult to test. Here, by treating 25km output from the GFDL-HiRAM-C360 model as “observations” for both past and future periods, we quantify the ability of three different statistical downscaling methods (season

AB - Climate change requires a global perspective to understand the past and explore the future. The impacts of climate change, however, are experienced mainly at the local to regional level. Downscaling techniques are commonly used to bridge the gap between the spatial scales at which climate is modeled vs. the scales at which impact assessments require climate projections. One of the most common assumptions made by downscaling is that of stationarity: that current-day relationships between climate variables, relationships that cannot be directly represented by fundamental physical equations but rather must be parameterized or statistically modeled, hold true under very different future conditions. As future observations are not yet available, the validity of this assumption is difficult to test. Here, by treating 25km output from the GFDL-HiRAM-C360 model as “observations” for both past and future periods, we quantify the ability of three different statistical downscaling methods (season

M3 - Paper

ER -