Breakout Summary Report

Breakout Description

The goal of this session was to bring together modelers and observationalists across all working groups to discuss approaches, progress, and challenges in (1) quantifying spatial heterogeneity in aerosol and cloud properties, (2) developing subgrid parameterizations, and (3) designing sampling strategies. This topic is relevant to ARM/ASR efforts as it is fundamental to the understanding of aerosol and cloud processes, especially aerosol-cloud interactions, and for connecting point observations and models across different scales. The relevance will only increase in future years with DOE/BER’s growing initiatives in urban and coastal science where spatial heterogeneity is a prominent feature. 

Main Discussion

We started this session with two 10-mins overview presentations. The first was by Jerome Fast from the aerosol perspective and the second from Zhibo Zhang from the cloud perspective.

We then split participants into small groups (about 10 people each) and conducted structured brainstorming exercises on the following questions:
Fundamentals

1. For which quantities and processes do we need to consider “heterogeneity” (drivers and impacts) from aerosol and cloud perspectives?


2. What do you think at what length scales does the heterogeneity have significant impacts and what are the related physical processes?

Observations

3. With current available ARM data or current/ongoing field campaign designs (e.g.,  ground-based, TBS, UAS, AAF), what heterogeneity-related problems can we tackle?

4. What observational design modifications to typical ARM deployments would enable better study of spatial heterogeneity and subgrid scale variability?

Modeling

5. For the effects of heterogeneity, what can be solved by high-resolution modeling (as computational power increases) and what cannot be solved?  

6. For those problems that can be solved by increasing resolution, is our model ready for it?  For those that cannot be solved, which directions should we pursue?   

After each block of questions, each group reported to the rest what they had discussed.

Key Findings

The two overview talks illustrated to what extent spatial heterogeneity matters for understanding aerosol-cloud interactions and gave several examples where we still have a long way to go. The small-group discussion was very lively and touched on many aspects of this complex topic. Below is a  brief summary of the main themes discussed in the presentations and for each of the questions.

1) A key quantity for which heterogeneity needs to be considered is CCN at different supersaturations. This, in turn, requires that we need to know aerosol number concentration, composition, size, and mixing state at the same spatial scales. At the same time, we need to know the spatial heterogeneity of clouds. Only when the two are collocated can there by “aerosol-cloud-interactions”.

2) Multiple length scales are relevant here. What matters is CCN entrained into clouds.  CCN outside of a cloud has no impact, but it is useful to understand the life cycle of CCN before activation. Updrafts can be 10s of m to a few km in diameter, so knowing CCN at those scales and if they are likely entrained into clouds is important. Clouds usually occur in large populations (organized or not), so having variability over 10s of km is needed.  Having information at cloud scale is useful for research, but climate models will be km scale at best for many years to come so knowing how to use the cloud-scale information for larger scale models is still important.

3) Several instruments need to be employed and coordinated to assemble a full picture. TBS is well suited for vertical profiles, to get CCN and aerosol property measurements up to cloud base. UAS would be good horizontally for cloud scale measurements and Bombardier for regional variations (flies too fast for smaller clouds but can carry a large payload including all the instruments that are necessary).  The EMSL post-analysis of aerosols can provide some info on mixing state, but limited sample size is a problem. The ability to get TBS, UAS, and Bombardier at the same time is also a logistical problem.  ARM usually has separate calls for each, so logistically they might not all be available at the same time. Ideally one would want to write one proposal for all and how the measurements would be coordinated instead of separate proposals.  A POPS-like network would be valuable on top of that because of the continuous measurements that can be made that won’t be possible from TBS, UAS, and Bombardier. Another idea would be to use a blimp which would have the advantage that it can carry a large payload.  

4) We are getting closer to be able to simulate aerosol #, composition, size distribution and mixing state; however, it is very expensive.  Obtaining high-resolution emission data is a challenge.

5) Models are – in principle – ready for everything except mixing state and obtaining suitable boundary conditions.  Seems that more work is needed on that (how detailed of a representation is needed to sufficiently represent CCN and how to simulate the processes efficiently?).  For example, WRF is not necessarily up to the task (too much baggage, older code, etc).

6) It is also valuable to measure aerosol emissions and cloud/PBL responses over different land cover/type uses, such as a network within a few tens of kilometers under similar meteorological conditions.  This is not only for land surface heterogeneity effect, but accounts for the upstream/downstream effects on the development of convective PBL and clouds.

7) Cloud ice retrievals are needed for cloud phase partitions and variabilities, especially those related to mixed-phase clouds. Especially during a cloud regime transition, such as cold air outbreaks, a lagrangian in situ sampling will be helpful to follow the air mass.

Issues

The room (lecture theater-style) wasn’t very well suited for the small group discussion format of this session. For example, we would have liked to use Post-it notes to document the small-group discussions, but this simply wasn’t practical in this setting.  If we do this in the future, we have to remember to check the room assignment beforehand.

Needs

The fact that we had many participants from both the cloud and aerosol working groups confirmed that this topic is highly relevant for a wide range of scientists within ARM/ASR. Not surprisingly, during the two-hour discussion, we were only able to scratch the surface of this complex topic. It became very clear that making progress relies on an unprecedented level of coordination between different observational platforms, but also between observationalists and modelers.

Decisions

N/A

Future Plans

This was the first time that we had conducted a session on this topic in this format with the small group discussion. We received positive feedback on the session format with the request to continue this conversation and to potentially elevate this to a workshop at some point in the future.

Action Items

We aim to continue the discussion at the next ARM/ASR PI meeting with the  aim to gauge support for a workshop on this topic.