Search Results

You are looking at 1 - 6 of 6 items for :

  • Regression analysis x
  • Sixth WMO Data Assimilation Symposium x
  • All content x
Clear All
Nicholas A. Gasperoni and Xuguang Wang

the variance of . To apply this method to the ensemble impact metric, a different regression is considered to be between the analysis in observation space and the forecast, such that for each observation l and state variable j , The use of the RCF method for the impact metric is inherently limited by the linear regression approximation needed to compute the group β s according to (6) , so there is an inherent limit to the forecast length at which it can be successfully applied. 3. Experiment

Full access
Juanzhen Sun, Hongli Wang, Wenxue Tong, Ying Zhang, Chung-Yi Lin, and Dongmei Xu

1. Introduction The variational data assimilation (DA) technique has been widely used in operational centers as well as in research communities to provide analysis and initialization for numerical models. The technique can be implemented with a three-dimensional (3DVar) or a four-dimensional (4DVar) variational data assimilation approach; the latter requires the use of a prediction model as the constraint. The variational method seeks to find the optimal analysis by minimizing a cost function

Full access
Daisuke Hotta, Tse-Chun Chen, Eugenia Kalnay, Yoichiro Ota, and Takemasa Miyoshi

preceded by small but statistically significant forecast degradations at shorter lead times. Kleist and Morgan (2005) pointed out another mechanism where 3) an analysis perturbation that improves a forecast does so by compensating for model errors rather than by reducing the true analysis error. Furthermore, 4) when working with forecast dropouts, it is difficult to rule out the possibility of “regression to the mean,” a statistical phenomenon that, in our context, can be stated as “for a forecast

Full access
Yicun Zhen and Fuqing Zhang

optimal ROI. Based on the above analysis, we propose the algorithm A0 ( Table 2 ) (while its results are shown in the left part of Table 3 ) for evaluating the optimal ROI. In algorithm A0, denotes the sample regression coefficient for the j th observation and i th grid point. Similarly is the regression coefficient computed by using the true covariance for the j th observation and i th grid point; d ji is the distance (in number of grid points) between the j th observation and i th grid

Full access
David Halpern, Dimitris Menemenlis, and Xiaochun Wang

-day and monthly data were 95% statistically higher at 140° than at 110°W, which, perhaps, was a result of the stronger EUC current at 140° than at 110°W in the presence of ubiquitous Kelvin and tropical instability waves. Regression analysis of the 3-day data shown in Fig. 2 had least squares lines of ECCO2 (m 2 s −1 ) = 0.68 ADCP (m 2 s −1 ) + 10.4 m 2 s −1 at 140°W and ECCO2 (m 2 s −1 ) = 0.55 ADCP (m 2 s −1 ) + 35.0 m 2 s −1 at 110°W. The 3-day ECCO2 values were higher (smaller) than

Full access
Robin J. T. Weber, Alberto Carrassi, and Francisco J. Doblas-Reyes

by calculating the overlap of the AI initial conditions and model probability distribution function (PDF) using the Bhattacharyya coefficient. As a useful reference, we carry out the same analysis for FFI and compare the performance. The paper is organized as follows. Section 2 formalizes AI and FFI. Section 3 connects AI to the mapping paradigm as introduced by Toth and Peña (2007) . Section 4 describes the experimental setup, along with the hierarchy of low-order models. Results and the

Full access