Stable Water Isotope Dynamics Can Constrain GCM Convective Processes during the MJO

Abstract:

This research aims to improve the representation of convective processes during the Madden-Julian oscillation (MJO) and other modes of intra-seasonal variability in the LMDZ atmospheric models, by making use of joint δD and H₂O (vapor) measurements.

In addition to atmospheric drying and wetting derived from the humidity measurements, the additional δD measurements provide enrichment and depletion information. This information is used to distinguish between different moistening and drying processes. For example, moistening due to ocean surface evaporation and due to rain re-evaporation can be distinguished, as re-evaporating moisture is more depleted in δD than surface evaporation.

We use mid-tropospheric IASI satellite δD and H₂O measurements to determine the humidity and δD evolution during about eight MJO events from 2010-2012 (including those monitored during the CINDY/DYNAMO campaign). Moreover, these evolutions are compared to the standard isotope enabled LMDZ GCM, as well as to sensitivity tests of key parameters (cold pool representation, precipitation efficiency, droplet size and fall speed, etc.) in the convection scheme.

The IASI measurements over the Indian ocean suggest that from 20 days to 5 days before the MJO peak, the main moisture source is oceanic surface evaporation, which is advected upwards by shallow convection. During the 5 days around the MJO peak, the moisture evolution is dominated by deep convection. Most inter-event variability occurs 5 to 10 days after the event, when 75% of the events are dominated by large scale condensation, while convection dominate the remaining quarter of the events. After this, the advection of relatively dry and enriched air brings back the state to the mean. Over the Maritime continent, similar δD-H₂O dynamics occur, but the variability of advected moisture dominates the inter-event variability.

The model captures the δD and H₂O dynamics of the MJO reasonably well. However, over the Indian ocean, the timing of the modeled depletion peak is too soon after the MJO peak. Initial results of sensitivity tests shows that parameters controlling rain re-evaporation rates have a significant influence on this timing of the δD minimum. Further sensitivity tests will be presented, together with the usefulness of δD and H₂O data to understand the MJO.