Climatological influence of Eurasian winter surface conditions on the Asian and Indo‐Pacific summer circulation in the NCEP CFSv2 seasonal reforecasts

Abstract

This study evaluates the possible influence of the winter surface conditions in Eurasia on the summer circulation over the Asian continent and Indo‐Pacific region. We have analysed multi‐seasonal ensemble reforecasts for 30 years (1979–2008) using the National Centers for Environmental Prediction Climate Forecast System version 2 initialized at the beginning of each month from January to May. It is found that the reforecasts initialized in winter (e.g., February) overestimate the snow cover fraction, depth and water equivalent, as well as surface albedo in the excessively snow‐covered portion of Eurasia from March to June. These biases are generated and perpetuated by a snow‐albedo feedback, leading to excessive upwards shortwave radiation reflected from the overly snow‐covered surface and an intense cold bias from the surface to mid‐troposphere. Originating over land, the cold bias is extended eastwards over the northwestern North Pacific by the advection of prevailing westerly winds. The cold air temperature in the broad mid‐latitude Asian‐Pacific region causes significantly lower geopotential heights at pressure levels in the middle and upper troposphere and thus increases the upper‐level westerly winds on its southern flank over the Asian continent and Indo‐Pacific. A slower than observed snow melting rate helps the winter cold bias persists well into the summer season in these runs. As a result, compared with the reforecasts initialized in spring (e.g., May), winter‐initialized reforecasts feature lower geopotential heights in the upper troposphere over Eurasia and a stronger subtropical jet over the Asian continent and the North Pacific from May to September, especially in early summer. The CFSv2 reforecasts in both sets of cases have too little total cloud fraction over Eurasia during June–August, leading to enhanced downwards shortwave radiation.

The Influence of Summer Deep Soil Temperature on Early Winter Snow Conditions in Eurasia in the NCEP CFSv2 Simulation

“…The magnitude of the annual cycle, however, is too large, with LST index values during peak wintertime being lower in the model simulation than CFSR values by~12°K and only 2-3°K lower during peak summer time. This cold LST bias in Eurasia in CFSv2, which recurs every year, is not due to the influence of initial shock (Shukla et al, 2017) or the initialization time (Shukla et al, 2019), but it is a permanent feature of the model. Based on observation products and reanalysis (example: Figure S1), stable snow cover in Eurasia is established in October, although some regions (e.g., northeastern Eurasia around 70°N) become snow covered in September.…”

“…It may be possible that the enhanced SCF anomalies in late fall can decrease the LST in the following winter. Our previous analysis (Shukla et al, 2019) of the CFSv2 seasonal reforecasts demonstrated that the reforecast runs initialized in February produce excessive snow cover and snow amount in Eurasia throughout the spring season due to an overactive snow-albedo feedback. As a result, the reforecast runs initialized in February produce colder temperatures at the surface and lower troposphere due to reflection of shortwave radiation over erroneously snow-covered areas of Eurasia during spring.…”

Cumulative Influence of Summer Subsurface Soil Temperature on North America Surface Temperature in the CFSv2

“…Despite the significant improvement in representing the physical processes that produce land surface properties in state‐of‐the‐art general circulation models, including snow, heat fluxes, land surface temperature (LST), soil moisture, and vegetation, a number of systematic biases and uncertainties of these properties persist (Douville, 2010; Dutra et al, 2011; Kim & Wang, 2007; Koster et al, 2004, 2011; Santanello et al, 2018; Seneviratne et al, 2006, 2010; van den Hurk et al, 2011, 2012). The inadequate representation of essential processes determining the propagation of information through the hydrological cycle in the general circulation models, as well as insufficient observations that are used to initialize land surface models, is a major cause for large bias in the mean climate and inaccurate evolution of interannual variations, which inevitably affects the prediction skill of air temperature, precipitation, and other surface properties (Delworth & Manabe, 1989; Douville, 2010; Guo et al, 2006; Koster et al, 2010, 2011; Koster & Suarez, 2003; Roesch, 2006; Roundy et al, 2014; Seneviratne et al, 2010; Shukla et al, 2019).…”

“…Many previous studies have used the National Centers for Environmental Prediction (NCEP) Coupled Forecast System version‐2 (CFSv2; Saha et al, 2014) to evaluate the seasonal prediction and simulation of El Niño–Southern Oscillation, Asian summer monsoon and other important atmospheric and oceanic processes in the reforecasts and long‐term simulations (Peng et al, 2012; Shukla & Huang, 2015; Dirmeyer & Halder, 2017; Shukla et al, 2017; Huang et al, 2017; Broxton et al, 2017; Shukla, Huang, Dirmeyer, & Kinter, 2019, Shukla, Huang, Dirmeyer, Kinter, Shin, & Marx, 2019, and papers cited therein). Using NCEP CFSv2 Reanalysis and Reforecast (CFSRR; Saha et al, 2014), He et al (2016, 2018) demonstrated that CFSv2 reproduces the spatial distribution of snow cover fraction (SCF) climatology in April at 0 lead month and snow water equivalent (SWE) during spring for a lead time of 1–3 months in Eurasia.…”

“…The excessive snow amount and its corresponding cold bias in LST also prolong the spring snow melting to early summer, which significantly influences the model‐simulated summer climate state in Eurasia. More recently, Shukla, Huang, Dirmeyer, Kinter, Shin, and Marx (2019) further showed that CFSv2 underestimates the amplitude of seasonal cycle in the deep soil temperature in 100–200 cm, resulting in a severe cold SUBT bias during summer in western Eurasia although its winter bias is relatively small. This systematic error of the summer SUBT can influence the subsequent seasonal transition and perpetuates the surface climate state significantly.…”