Correction of monthly SST forecast in CFSv2 using the Local Dynamical Analog method

View More View Less
  • 1 Frontiers Science Center for Deep Ocean Multispheres and Earth System (FDOMES)/Key Laboratory of Physical Oceanography/Institute for Advanced Ocean Studies, Ocean University of China, Qingdao 266100, China.
  • 2 Laboratory for Ocean Dynamics and Climate, Pilot Qingdao National Laboratory for Marine Science and Technology, Qingdao 266237, China.
© Get Permissions
Restricted access

Abstract

Numerical seasonal forecasts in Earth science always contain forecast errors that cannot be eliminated by improving the ability of the numerical model. Therefore, correction of model forecast results is required. Analog-correction is an effective way to reduce model forecast errors, but the key question is how to locate analogs. In this paper, we updated the Local Dynamical Analog (LDA) algorithm to find analogs and depicted the process of model error correction as the LDA-correction scheme. The LDA-correction scheme was firstly applied to correct the operational seasonal forecasts of sea surface temperature (SST) over the period 1982–2018 from the state-of-the-art coupled climate model named NCEP Climate Forecast System version 2.

The results demonstrated that the LDA-correction scheme improves forecast skill in many regions as measured by the correlation coefficient and Root Mean Square Error, especially over the extratropical eastern Pacific and tropical Pacific, where the model has high simulation ability. El Niño-Southern Oscillation (ENSO) as the focused physics process is also improved. The seasonal predictability barrier of ENSO is in remission and the forecast skill of Central Pacific ENSO also increases due to the LDA-correction method. The intensity of ENSO mature phases is improved. Meanwhile, the ensemble forecast results are corrected, which proves the positive influence from this LDA-correction scheme on the probability forecast of cold and warm events. Overall, the LDA-correction scheme, combining statistical and model dynamical information, is demonstrated to be readily integrable with other advanced operational models and has the capability to improve forecast results.

Corresponding author: Jianping Li (ljp@ouc.edu.cn)

Abstract

Numerical seasonal forecasts in Earth science always contain forecast errors that cannot be eliminated by improving the ability of the numerical model. Therefore, correction of model forecast results is required. Analog-correction is an effective way to reduce model forecast errors, but the key question is how to locate analogs. In this paper, we updated the Local Dynamical Analog (LDA) algorithm to find analogs and depicted the process of model error correction as the LDA-correction scheme. The LDA-correction scheme was firstly applied to correct the operational seasonal forecasts of sea surface temperature (SST) over the period 1982–2018 from the state-of-the-art coupled climate model named NCEP Climate Forecast System version 2.

The results demonstrated that the LDA-correction scheme improves forecast skill in many regions as measured by the correlation coefficient and Root Mean Square Error, especially over the extratropical eastern Pacific and tropical Pacific, where the model has high simulation ability. El Niño-Southern Oscillation (ENSO) as the focused physics process is also improved. The seasonal predictability barrier of ENSO is in remission and the forecast skill of Central Pacific ENSO also increases due to the LDA-correction method. The intensity of ENSO mature phases is improved. Meanwhile, the ensemble forecast results are corrected, which proves the positive influence from this LDA-correction scheme on the probability forecast of cold and warm events. Overall, the LDA-correction scheme, combining statistical and model dynamical information, is demonstrated to be readily integrable with other advanced operational models and has the capability to improve forecast results.

Corresponding author: Jianping Li (ljp@ouc.edu.cn)
Save