1B.4 A seamless assessment of the role of convection in the water cycle of the West African Monsoon

Monday, 31 March 2014: 9:00 AM
Pacific Salon 4 & 5 (Town and Country Resort )
Cathryn E. Birch, UK Met Office, Leeds, West Yorkshire, United Kingdom; and D. J. Parker, J. H. Marsham, D. Copsey, and L. Garcia-Carreras

A seamless suite of 40-day UK Met Office Unified Model simulations over West Africa during summer 2006 are analysed to investigate the causes of biases in the position of the rainbelt and to understand the role of convection in the regional water budget. The simulations include climate, global operational and limited-area runs (grid-spacings from 1.5 to 40km), including two 12-km runs, one with parameterised and one with explicit convection.

The most significant errors in the water-cycle terms occur in the simulations with parameterised convection, associated with the diurnal cycle and the location of the convection. Errors in the diurnal cycle increase the northward advection of moisture out of the Sahel towards the Sahara, but decrease the advection of moisture into the Sahel from further south, which limits the availability of moisture for Sahelian rainfall. These biases occur within the first 24 hours, showing that they originate from the representation of fast physical processes, specifically, the convection scheme. Once these rainfall regimes have been established, the terms of the water budgets act to reinforce the biases, effectively locking the rainbelt's latitude.

One of the simulations with parameterised convection does, however, produce a better latitudinal distribution of rainfall because on the first day it is better able to trigger convection in the Sahel. Accurate representation of the diurnal cycle of convection and the ability to trigger convection in a high convective inhibition environment is key to capturing the water cycle of the region and will improve the representation of the West African Monsoon.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner