Application of blocking diagnosis methods to General Circulation Models. Part I: a novel detection scheme

Abstract

This paper aims to provide a new blocking definition with applicability to observations and model simulations. An updated review of previous blocking detection indices is provided and some of their implications and caveats discussed. A novel blocking index is proposed by reconciling two traditional approaches based on anomaly and absolute flows. Blocks are considered from a complementary perspective as a signature in the anomalous height field capable of reversing the meridional jet-based height gradient in the total flow. The method succeeds in identifying 2-D persistent anomalies associated to a weather regime in the total flow with blockage of the westerlies. The new index accounts for the duration, intensity, extension, propagation, and spatial structure of a blocking event. In spite of its increased complexity, the detection efficiency of the method is improved without hampering the computational time. Furthermore, some misleading identification problems and artificial assumptions resulting from previous single blocking indices are avoided with the new approach. The characteristics of blocking for 40 years of reanalysis (1950–1989) over the Northern Hemisphere are described from the perspective of the new definition and compared to those resulting from two standard blocking indices and different critical thresholds. As compared to single approaches, the novel index shows a better agreement with reported proxies of blocking activity, namely climatological regions of simultaneous wave amplification and maximum band-pass filtered height standard deviation. An additional asset of the method is its adaptability to different data sets. As critical thresholds are specific of the data set employed, the method is useful for observations and model simulations of different resolutions, temporal lengths and time variant basic states, optimizing its value as a tool for model validation. Special attention has been paid on the devise of an objective scheme easily applicable to General Circulation Models where observational thresholds may be unsuitable due to the presence of model bias. Part II of this study deals with a specific implementation of this novel method to simulations of the ECHO-G global climate model.

Appendix: Computation of the anomaly field

Assuming that the length of the data set is K years, N the number of data points per year and M the number of data points per month so that 1 ≤ t ≤ K N and 1 ≤ τ ≤ N, a four-step procedure has been applied: (1) evaluation of a running annual mean with special treatment of the ends of the time series (Eq. 5); (2) determination of a running monthly mean of the anomaly relative to the annual mean (Eq. 6); (3) computation of the mean seasonal cycle (Eq. 7); (4) computation of the anomaly field by extracting the annual mean and the seasonal cycle from the total flow (Eq. 8).

$$ \overline{z} (t) = \left\{ {\begin{array}{*{20}l} {{\frac{1}{N + 1}}\sum\limits_{{k = t - {N / 2}}}^{{k = t + {N / 2}}} {z\left( k \right)} } \hfill & {{N/ 2} < t < KN - {N / 2}} \hfill \\ {\overline{z} \left( {{N / 2} + 1} \right)} \hfill & {1 \le t \le {N / 2}} \hfill \\ {\overline{z} \left( {KN - {N / 2} - 1} \right)} \hfill & {K N - {N / 2} \le t \le K N} \hfill \\ \end{array} } \right. $$

(5)

$$ z*\left( t \right) = \left\{ {\begin{array}{*{20}l} {{\frac{1}{M + 1}}\sum\limits_{{k = t - M /2}}^{{k = K N - M/2}} {\left( {z\left( k \right) - \overline{z} \left( k \right)} \right)} } \hfill & {M/ 2 < t < K N - M/2} \hfill \\ {z*\left( {{M /2} + 1} \right)} \hfill & {1 \le t \le {M/2} } \hfill \\ {z*\left( {K N - {M / 2} - 1} \right)} \hfill & {K N - {M/ 2} \le t \le K N} \hfill \\ \end{array} } \right. $$

(6)

$${\hat{z}} ( \tau ) = {\frac{1}{K}}\sum\limits_{k = 1}^{k = K} {z*( N (k-1) + \tau)}\, {1 \le \tau \le N} $$

(7)

$$ {z'\left( t \right) = z\left( t \right) - \overline{z} \left( t \right) - \hat{z} \left( \tau \right)} \quad {1 \le \tau \le N} \quad {1 \le t \le K N} $$

(8)