An Investigation of the Time Scales of Development of the Double-ITCZ Bias

Abstract

The predominant precipitation feature in the deep tropics is a band of convective precipitation known as the Intertropical Convergence Zone (ITCZ). Many global climate models simulate a secondary band of excessive precipitation south of the equator over the east and central Pacific Ocean. This study investigates this bias by examining the time scales associated with shifts in the ITCZ when the model is forced both locally and remotely. A simple two box model is used to demonstrate a differentiation in temperature response time scale between locally and remotely forced regions. When two heat reservoir are tightly coupled, and a steady perturbation heat flux is applied to one box, the temperature response of the unforced box is delayed as compared to the temperature response in the forced box provided the coupling between the boxes is sufficiently weak. This result is extended to a more realistic model by abruptly adding a steady anti-symmetric heat flux convergence to the slab ocean heat flux convergence of the Geophysical Fluid Dynamics Laboratory's Atmospheric Model version 2.1. The tropical ITCZ response time scale of the model is characterized using the response time scale of the difference in northern and southern hemispheric mean tropical surface temperature. The response time scale of this quantity is shown to scale linearly with the model's combined ocean-atmosphere heat capacity. When the perturbation is added in the mid-latitudes, the resultant ITCZ shift is amplified by subtropical clear-sky radiation and cloud radiative feedbacks. For a realistic mixed layer depth of 50 m, the tropics reach their new equilibrium in just under eight years. When a perturbation of similar strength is applied in the tropics, the response time scale is much faster with the model reaching equilibrium in nearly half the time required to equilibrate to the mid-latitude perturbation. The final portion of this thesis investigates the transient development of the double-ITCZ bias in the National Center for Atmospheric Research's fully coupled Community Earth System Model version 1.2. The model is shown to develop a double ITCZ bias in the east Pacific within the first year of integration with little sensitivity to the convective parameterization employed. The bias is fully developed by the end of the second year of integration. The rapid onset of the bias suggests either a local source or strong coupled feedbacks are responsible for the development of the bias. The double-ITCZ bias is also associated with an overly cool equatorial Pacific cold tongue. The central Pacific precipitation bias is lessened when the deep convective parameterization is disabled.