Temporal evolution of hydroclimatic teleconnection and a time-varying model for long-lead prediction of Indian summer monsoon rainfall

Abstract

Several cases of failure in the prediction of Indian Summer Monsoon Rainfall (ISMR) are the major concern for long-lead prediction. We propose that this is due to the temporal evolution of association/linkage (inherent concept of temporal networks) with various factors and climatic indices across the globe, such as El Niño-Southern Oscillation (ENSO), Equatorial Indian Ocean Oscillation (EQUINOO), Atlantic Multidecadal Oscillation (AMO), North Atlantic Oscillation (NAO), Pacific Decadal Oscillation (PDO) etc. Static models establish time-invariant (permanent) connections between such indices (predictors) and predictand (ISMR), whereas we hypothesize that such systems are temporally varying in nature. Considering hydroclimatic teleconnection with two major climate indices, ENSO and EQUINOO, we showed that the temporal persistence of the association is as low as three years. As an application of this concept, a statistical time-varying model is developed and the prediction performance is compared against its static counterpart (time-invariant model). The proposed approach is able to capture the ISMR anomalies and successfully predicts the severe drought years too. Specifically, 64% more accurate performance (in terms of RMSE) is achievable by the recommended time-varying approach as compared to existing time-invariant concepts.

Contrasting features of hydroclimatic teleconnections and the predictability of seasonal rainfall over east and west Japan

“…The threshold of the EED is 3.84 (at a 5% significance level with 1 degree of freedom), so the edges for which the EED < 3.84 are to be excluded. To F I G U R E 5 (a) 1-6 months lagged global fields of the sea surface temperature (SST) anomaly difference obtained by subtracting the mean global field associated with above-normal rainfall events from the mean global field associated with below-normal rainfall events in December over east Japan; (b) same as for (a), but in January over east Japan; and (c) same as for (a), but in February over east Japan check the acceptability of the obtained graph structure at a particular confidence level, a test statistic, known as the deviance, can be used (Dutta and Maity, 2018).…”

“…The prediction model is developed using C-Vine copula approach, in which a sequence of trees is identified to develop the conditional distribution of the target variable given the parents (Xiao, 2011;Bauer et al, 2012;Liu et al, 2015;Righi et al, 2015;Dalla Valle et al, 2016). The selection of each tree is based on a maximum spanning tree algorithm, where edge weights are chosen to reflect the dependencies and the final tree can be used for the prediction of the target variable given the input variables (Dutta and Maity, 2018).…”

“…For instance, considering the regions of China, Australia, Latin America and India, various local and global climatic forcings (e.g. the ENSO, AMO, PDO, NAO and IOD) impact rainfall at various spatio-temporal scales (Gu et al, 2009;Wu and Zhang, 2010;Fu et al, 2013;King et al, 2014;Ye, 2014;Cid-Serrano et al, 2015;Dutta and Maity, 2018;Nair et al, 2018).…”

Climate and rivers

“…Srivastava, Pradhan, Goswami, and Rao () in noting the changing relationships between major modes of climate variability and the IM, point to the shift in the tropical climate in the late 1970s, especially the warming of the central Pacific and the Indian Ocean (IO), as the driver of this nonstationarity. Although not focusing of intraseasonal to interannual river flow forecasting, the work of Dutta and Maity () bears implications for attempts to develop river flow prediction models for the IM region using teleconnection indices such as ENSO as the predictors. They found that a statistical time‐varying prediction model for IM rainfall, which accounted for nonstationarity in ENSO/IOD/IM associations, to be superior in performance compared with a time‐invariant model that assumed stationarity.…”

Spatial variation in long‐lead predictability of summer monsoon rainfall using a time‐varying model and global climatic indices

“…Another important aspect is that the summer monsoon rainfall at regional scale over the Indian domain responds to the above‐mentioned large‐scale climatic indices in complex ways. Dutta and Maity (2018) studied the time‐varying association among two large‐scale climatic indices and ISMR and concluded the need to carry out a detailed analysis at finer spatial scale with a larger pool of large‐scale climatic indices. Several other studies have been performed on Indian rainfall; however, it is difficult to treat India as a single unit for interpreting the association with the large‐scale indices, as the association have seasonal and regional differences (Maity and Nagesh, 2006; Vathsala and Koolagudi, 2017).…”