The effect of increased entrainment on monsoon precipitation biases in a GCM Stephanie Bush (University of Reading/UK Met Office), Andrew Turner (Reading), Steve Woolnough (Reading), Gill Martin (Met Office) WGNE April 2013

JWCRP Stephanie Bush University of Reading Motivation  The south Asian monsoon is accompanied by a dramatic shift in the ITCZ that brings most of India and SE Asia’s annual rainfall  Most GCMs have significant, consistent biases in monsoon rainfall, including the UK Met Office Unified Model (e.g. Sperber et al. 2013)  N96 (~200 km) L85, 20 year AMIP, HadGEM3 Global Atmosphere 3.0 (Walters et al. 2011)  Convection parameterized with three components: deep, mid-level and shallow (extensively modified version of Gregory &Rowntree 1990)  Monsoon biases are highly sensitive to convective parameters (e.g. Bechtold et al. 2008) Bias: GPCP and ERA Interim UM JJAS precipitation and 850 hPa circulation Mid-level and Deep Entrainment increased by a factor of 1.5  Increasing the convective entrainment rate tends to improve intraseasonal variability such as the MJO (e.g. Klingaman et al. submitted, Hirons et al. 2012, Del Genio et al. 2012)  Increasing the convective entrainment rate globally in the model decreases several long standing monsoon precipitation biases while increasing others (Kim et al. 2011) :  Improves rainfall biases over Western Equatorial Indian Ocean (WEIO) and steep continental orography  Slightly improves rainfall over India  Worsens rainfall biases over Western North Pacific (WNP) and Maritime Continent Bush et al in preparation

JWCRP Stephanie Bush University of Reading Increasing the entrainment rate Increasing the entrainment rate increases mixing of moist convective plumes with drier air aloft. Less entrainmentMore entrainment Convective Plume Entraining environmental air & detraining air from plume Increasing humidity Surface With increased entrainment we expect: Suppressed convection, decreased precipitation, moistening of column. Stronger effects in drier environments.

JWCRP Stephanie Bush University of Reading Targeted Experiments Bush et al in preparation UM JJAS precipitation and 850 hPa circulation Entrainment globally increased by a factor of 1.5 Use experiments targeted at land, WEIO and WNP to: Analyze the local response to entrainment change Determine how modifying specific rainfall biases affects other rainfall biases Inform efforts to reduce biases Use experiments targeted at land, WEIO and WNP to: Analyze the local response to entrainment change Determine how modifying specific rainfall biases affects other rainfall biases Inform efforts to reduce biases

JWCRP Stephanie Bush University of Reading Targeted Experiments Entrainment increased only over WEIO Entrainment increased only over land Entrainment increased only over WNP Bush et al in preparation UM JJAS precipitation and 850 hPa circulation Entrainment globally increased by a factor of 1.5

JWCRP Stephanie Bush University of Reading Targeted Experiments Entrainment increased only over WEIO Entrainment increased only over land Entrainment increased only over WNP Bush et al in preparation Entrainment globally increased by a factor of 1.5 The global response can be (mostly) understood as a combination of the local responses. Still need to understand why the regions locally respond so differently to one another. Decreased precipitation in other regions is not responsible for the enhanced precipitation in the WNP The global response can be (mostly) understood as a combination of the local responses. Still need to understand why the regions locally respond so differently to one another. Decreased precipitation in other regions is not responsible for the enhanced precipitation in the WNP Diagnostics July 1 initialized ensemble Dynamical spin-up Moisture budget Time-step rainfall rates Relationship between precipitation and moisture

JWCRP Stephanie Bush University of Reading Spin-up experiments Increased entrainment imposed on July 1: after monsoon circulation established Bush et al in preparation WEIO experimentWNP experiment Global experiment Initial response: Decreased precipitation Gill circulation response to decreased latent heating (Gill 1980)

JWCRP Stephanie Bush University of Reading Moisture Budget Time Series WNP Precipitation (mm/day)  Control x Increased Entrainment Bush et al in preparation WEIO WNP Precipitation (mm/day)Horizontal Moisture Flux (mm/day) Evaporation (mm/day)Column Water Vapor (mm)

JWCRP Stephanie Bush University of Reading Precipitation vs. CWV Precipitation (mm/day) Column Water Vapor (mm) Daily grid-point values in tropical (30 N – 30 S) regions of ascent (Ω< 0) Daily grid-point CWV values in: WEIO WNP Increasing the entrainment rate increases the sensitivity of precipitation to moisture and, on average, decreases precipitation at lower CWV and increases it at higher CWV. Different distributions of moisture in the WEIO and WNP may affect their response to increased entrainment. Control Globally Increased Entrainment Control Globally Increased Entrainment

JWCRP Stephanie Bush University of Reading Conclusions Mid level & deep entrainment increased by a factor of 1.5 Increasing the convective entrainment rate has a large effect on the monsoon mean state Local responses are consistent with global response, but feedbacks strengthen them Can have a very different mean state response in similar regions. Why do the WEIO and WNP respond differently? Possible factors: Difference in seasonal cycle Difference in initial moisture content Difference in moisture available to advect into a region Equatorial vs. off equatorial dynamical response Future work: Aquaplanet and SCM experiments UM JJAS precipitation and 850 hPa circulation

#15 - Aerosol impact on the South Asian Summer monsoon precipitation simulated by CMIP5 historical experiment (Liang Guo et al.) Over South Asia, precipitation changes resemble change with aerosol forcing only. This resemblance exists both in temporal evolution and spacial pattern over South Asia.

#21- Model error: the Arabian Sea winter cold bias (Marathayil et al., 2013, ERL) Many CMIP models exhibit cold SSTs in the northern Arabian Sea during winter and spring. These link a series of coupled biases in the Indian Ocean, ultimately related to cold temperatures over land further north. Composite of 7 coldest minus 7 warmest models based on northern Arabian Sea SST Scatter of northern Arabian Sea SST biases versus meridional wind biases (DJF)

#22 - The role of the Indian peninsula in the South Asian monsoon (Andrew Turner et al.) Role of Indian peninsula tested by: removing whole peninsula; removing Western Ghats only; using 100% bare soil or 100% lake at the surface. Change due to 100% bare soil or 100% lake no-peninsular mean climate and change No-Ghats mean climate and change

© Crown copyright Met Office Thursday #36: The influence of dynamic vegetation on the simulation of monsoons in the HadGEM2 family. Gill Martin and Richard Levine  Persistent tropical rainfall biases result in a large increase in bare soil in HadGEM2-ES.  This affects the rainfall through both the surface characteristics and increased atmospheric dust. Daily means Bare soil fraction in HadGEM2-ES Clear-sky incoming SW at surface Change due to vegetation Change due to vegetation Change due to dust (no dust effect)radiative effects Change due to vegetation Change due to dust