Final Report - From Measurements to Models: Cross-Comparison of Measured and Simulated Behavioral States of the Atmosphere

Del Genio, Anthony D ; Hoffman, Forrest M ; Hargrove, Jr, William W

NASA Goddard Institute for Space Studies

关键词: Regimes;    Clustering;    Clouds;    Climate;    54 Environmental Sciences Clouds;   

DOI  :  10.2172/918047
RP-ID  :  DOE/ER/63958
RP-ID  :  AI02-05ER63958
RP-ID  :  918047

美国|英语

来源: UNT Digital Library

【 摘 要 】

The ARM sites and the ARM Mobile Facility (AMF) were constructed to make measurements of the atmosphere and radiation system in order to quantify deficiencies in the simulation of clouds within models and to make improvements in those models. While the measurement infrastructure of ARM is well-developed and a model parameterization testbed capability has been established, additional effort is needed to develop statistical techniques which permit the comparison of simulation output from atmospheric models with actual measurements. Our project establishes a new methodology for objectively comparing ARM measurements to the outputs of leading global climate models and reanalysis data. The quantitative basis for this comparison is provided by a statistical procedure which establishes an exhaustive set of mutually-exclusive, recurring states of the atmosphere from sets of multivariate atmospheric and cloud conditions, and then classifies multivariate measurements or simulation outputs into those states. Whether measurements and models classify the atmosphere into the same states at specific locations through time provides an unequivocal comparison result. Times and locations in both geographic and state space of model-measurement agreement and disagreement will suggest directions for the collection of additional measurements at existing sites, provide insight into the global representativeness of the current ARM sites (suggesting locations and times for use of the AMF), and provide a basis for improvement of models. Two different analyses were conducted: One, using the Parallel Climate Model, focused on an IPCC climate change scenario and clusters that characterize long-term changes in the hydrologic cycle. The other, using the GISS Model E GCM and the ARM Active Remotely Sensed Cloud Layers product, explored current climate cloud regimes in the Tropical West Pacific.