Effects of explicit convection on global land‐atmosphere coupling in the superparameterized CAM

Abstract

Conventional global climate models are prone to producing unrealistic land‐atmosphere coupling signals. Cumulus and convection parameterizations are natural culprits but the effect of bypassing them with explicitly resolved convection on global land‐atmosphere coupling dynamics has not been explored systematically. We apply a suite of modern land‐atmosphere coupling diagnostics to isolate the effect of cloud Superparameterization in the Community Atmosphere Model (SPCAM) v3.5, focusing on both the terrestrial segment (i.e., soil moisture and surface turbulent fluxes interaction) and atmospheric segment (i.e., surface turbulent fluxes and precipitation interaction) in the water pathway of the land‐atmosphere feedback loop. At daily timescales, SPCAM produces stronger uncoupled terrestrial signals (negative sign) over tropical rainforests in wet seasons, reduces the terrestrial coupling strength in the Central Great Plain in American, and reverses the coupling sign (from negative to positive) over India in the boreal summer season—all favorable improvements relative to reanalysis‐forced land modeling. Analysis of the triggering feedback strength (TFS) and amplification feedback strength (AFS) shows that SPCAM favorably reproduces the observed geographic patterns of these indices over North America, with the probability of afternoon precipitation enhanced by high evaporative fraction along the eastern United States and Mexico, while conventional CAM does not capture this signal. We introduce a new diagnostic called the Planetary Boundary Layer (PBL) Feedback Strength (PFS), which reveals that SPCAM exhibits a tight connection between the responses of the lifting condensation level, the PBL height, and the rainfall triggering to surface turbulent fluxes; a triggering disconnect is found in CAM.

Could Machine Learning Break the Convection Parameterization Deadlock?

“…To alleviate this problem, an interesting approach has been to use cloud “superparameterization (SP),” which computes the subgrid vertical heating and moistening profiles within a GCM grid cell by sampling a curtain of an embedded 2‐D CRM that uses convective permitting resolution (Grabowski, ; Khairoutdinov et al, ). This has led to many successes such as the possibility to rectify the diurnal continental cycle, to improve the representation of the MJO, and to represent both some MCS propagation and some degree of aggregation, and reduce overly strong land‐atmosphere coupling (Benedict & Randall, 2009; Grabowski, ; Holloway et al, , ; Khairoutdinov et al, ; Kooperman et al, , ; Pritchard & Somerville, ; Pritchard et al, ; Qin et al, ; Randall, ; Sun & Pritchard, ).…”

Global Effects of Superparameterization on Hydrothermal Land‐Atmosphere Coupling on Multiple Timescales

“…Recently, the first detailed analysis of SP from the perspective of modern coupling metrics showed that it favorably reduces hydrologic L‐A coupling on subdaily timescales (Sun & Pritchard, , SP16 hereafter). But three limitations of SP16 are (1) the use of diagnostic correlative metrics of coupling; (2) the focus on short (subdaily) timescales of L‐A coupling; and (3) the focus on hydrologic L‐A coupling alone.…”

Evaluating Soil Moisture Feedback on Convective Triggering: Roles of Convective and Land‐Model Parameterizations

“…A wellknown approach is to perform soil moisture perturbation experiments in climate models (Koster et al, 2003(Koster et al, , 2006 or simplified boundary-layer models (Ek & Holtslag, 2004;Findell & Eltahir, 2003;Gentine et al, 2013). Such controlled experiments are effective in identifying causal relationships, but are subject to errors in model representation of convective clouds (Chen et al, 2017;Sun & Pritchard, 2016), boundary-layer turbulence (Santanello et al, 2011), and ET (I. N. Williams et al, 2016).…”