The VAMOS Ocean-Cloud-Atmosphere-Land Study Regional Experiment (VOCALS-REx): goals, platforms, and field operations

The VAMOS1 Ocean-Cloud-Atmosphere-Land Study Regional Experiment (VOCALS-REx) was an international field program designed to make observations of poorly understood but critical components of the coupled climate system of the southeast Pacific. This region is characterized by strong coastal upwelling, the coolest SSTs in the tropiCorrespondence to: R. Wood (robwood@atmos.washington.edu) 1Variability of the American Monsoon Systems, an international CLIVAR program. cal belt, and is home to the largest subtropical stratocumulus deck on Earth. The field intensive phase of VOCALSREx took place during October and November 2008 and constitutes a critical part of a broader CLIVAR program (VOCALS) designed to develop and promote scientific activities leading to improved understanding, model simulations, and predictions of the southeastern Pacific (SEP) coupled ocean-atmosphere-land system, on diurnal to interannual timescales. The other major components of VOCALS are a modeling program with a model hierarchy ranging from the local to global scales, and a suite of extended observations from regular research cruises, instrumented moorings, Published by Copernicus Publications on behalf of the European Geosciences Union. 628 R. Wood et al.: VOCALS operations and satellites. The two central themes of VOCALS-REx focus upon (a) links between aerosols, clouds and precipitation and their impacts on marine stratocumulus radiative properties, and (b) physical and chemical couplings between the upper ocean and the lower atmosphere, including the role that mesoscale ocean eddies play. A set of hypotheses designed to be tested with the combined field, monitoring and modeling work in VOCALS is presented here. A further goal of VOCALS-REx is to provide datasets for the evaluation and improvement of large-scale numerical models. VOCALSREx involved five research aircraft, two ships and two surface sites in northern Chile. We describe the instrument payloads and key mission strategies for these platforms and give a summary of the missions conducted.

and satellites. The two central themes of VOCALS-REx focus upon (a) links between aerosols, clouds and precipitation and their impacts on marine stratocumulus radiative properties, and (b) physical and chemical couplings between the upper ocean and the lower atmosphere, including the role that mesoscale ocean eddies play. A set of hypotheses designed to be tested with the combined field, monitoring and modeling work in VOCALS is presented here. A further goal of VOCALS-REx is to provide datasets for the evaluation and improvement of large-scale numerical models. VOCALS-REx involved five research aircraft, two ships and two surface sites in northern Chile. We describe the instrument payloads and key mission strategies for these platforms and give a summary of the missions conducted.

Scientific motivation
Interactions between the South American continent and the Southeast Pacific (SEP) Ocean are extremely important for both the regional and global climate system. Figure 1 indicates some of the key features associated with these interactions. The great height and continuity of the Andes Cordillera forms a sharp barrier to zonal flow, resulting in strong winds (coastal jet) parallel to the coasts of Chile and Peru (Garreaud and Muñoz, 2005). This, in turn, drives intense oceanic upwelling along these coasts, bringing cold, deep, nutrient/biota rich waters to the surface. As a result, the coastal SEP sea-surface temperatures (SSTs) are colder along the Chilean and Peruvian coasts than at any comparable latitude elsewhere. The cold surface, in combination with warm, dry air aloft, is ideal for the formation of marine stratocumulus clouds, and supports the largest and most persistent subtropical stratocumulus deck in the world (Klein and Hartmann, 1993). The presence of this cloud deck has a major impact upon the Earth's radiation budget by reflecting solar radiation. This helps maintain cool SSTs, resulting in tight couplings between the upper ocean and lower atmosphere in this region. The unique climate of the SEP has been very sparsely observed, yet has great economic impact, with fishing in the Humboldt Current system representing 18-20% of the worldwide marine fish catch (source: UN LME report).
It is a challenge for global and regional models to successfully simulate the SE Pacific climate system, because of its sharp horizontal and vertical gradients and the importance of subgridscale and poorly resolve physical processes. Most coupled GCMs obtain SSTs that are too warm and have too few clouds over the SEP, and show unrealistic features in the simulation of the warm tropics downstream (deSzoeke and Xie, 2008). There are major uncertainties in the representation of key physical processes in these models, which may be contributing to these errors (e.g. Mechoso et al., 1995;Ma  et al., 1996). The representation of stratocumulus in large scale models over the SEP is improving in some global models, but most models continue to have large biases in the location and albedo of cloud and the boundary layer vertical structure (Bretherton et al., 2004;Wyant et al., 2010). Observations are highlighting the importance of drizzle precipitation to SEP marine stratocumulus (e.g. Bretherton et al., 2004;Caldwell et al., 2005;Comstock et al., 2005), and observations and models point to a significant role for drizzle in affecting stratocumulus cloud cover and radiative properties, in particular in promoting transitions from closed to open mesoscale cellular convection (e.g. Comstock et al., 2007;Savic-Jovcic and Stevens, 2008;Wang and Feingold, 2009;Wang et al., 2010) and the formation of so-called "pockets of open cells" (POCs) (Bretherton et al., 2004;Stevens et al., 2005). Physical parameterizations currently used in large scale models do not yet attempt to represent mesoscale interactions between precipitation and cloud cover.
There is evidence that precipitation in marine stratocumulus may be influenced by anthropogenic aerosols (e.g. Geoffroy et al., 2008;Brenguier and Wood, 2009), which suggests a potential role for aerosols to influence cloud macrostructure in addition to their microphysics. Aerosol indirect effects on warm clouds remain poorly treated in large scale numerical models (e.g. Lohmann and Feichter, 2005), chiefly because the overall impact of aerosols on cloud radiative properties depends upon numerous complex small scale and mesoscale dynamical responses which result in macrophysical cloud changes (Stevens and Feingold, 2009). Satellite and research cruise data show strong gradients in aerosol and cloud microphysical properties between the near-coastal and more remote marine region of the SEP , making this a region where the Twomey effect may be particularly strong (see e.g. George and Wood, 2010). Since this is also a region where clouds are prone to drizzle (Bretherton et al., 2004;Leon et al., 2008), it is also a region potentially wellsuited to the study of aerosol-cloud interactions.
In the SEP region there are important contributions to the atmospheric aerosol from both natural and anthropogenic sources (Tomlinson et al., 2007;Hawkins et al., 2010). Cloud droplet effective radii are small off the coast of Northern Chile, implying elevated concentrations of cloud droplets George and Wood, 2010;Painemal and Zuidema, 2010). These elevated concentrations are broadly downwind of major copper smelters whose combined sulfur emissions total approximately 1 TgS yr −1 , comparable to the entire sulfur emissions from large industrialized nations such as Mexico and Germany. Offshore transport events have been shown to lead to elevated droplet concentrations offshore (Huneeus et al., 2006). However, little is actually known about the aerosol composition in the region since there have been very few measurements. We do not yet know the extent of the anthropogenic influence, nor do we fully understand the complex chemistry occurring in the pristine boundary layer further offshore.
In the absence of cloud macrophysical responses, the reduced droplet effective radii resulting from increased concentrations of cloud droplets would increase the reflected solar radiation, and estimates of the component of the reflected shortwave radiation due to geographic variability in effective radius alone are ∼10-20 W m −2 or 20-40% of the mean reflected shortwave . The magnitude of these estimates is such that the indirect effects of aerosols on clouds could lead to significant decreases in the amount of solar radiation entering the ocean, with significant implications for the ocean heat budget. However, we do not yet fully understand the controls on cloud droplet concentration in the MBL, and it is possible that meteorological controls (e.g. precipitation sinks) in addition to aerosol sources may play a significant role. Further, we are beginning to understand that cloud responses to aerosols are not solely due to the Twomey effect alone, and that fast feedbacks can both enhance and counteract the Twomey effect (e.g. Ackerman et al., 2004;Xue et al., 2008).
Early estimates of surface heat fluxes from climatologies and numerical weather prediction models showed diverse conclusions as to whether or not the offshore ocean gained from or lost heat to the atmosphere. Observations from deployment of the IMET surface mooring beginning in 2000 near the location of the annual maximum in stratus cloud cover showed that the ocean gains about 40 W m −2 annually and was subject to over 1 m in evaporation (Colbo and Weller, 2007;deSzoeke et al., 2010). This surface forcing was applied to a relatively thin ocean surface mixed layer (annual maximum thickness of about 150 m) that lies over a relatively cold, fresh water mass formed to the south. For oceanographers, the challenge is to understand how the shallow ocean surface layer under the clouds maintains its temperature and salinity under this surface forcing.
Studies of the upper ocean heat budget offshore of the coastal upwelling zone indicate weak mean advection, energetic eddies, and the need for a source of relatively cold, fresh water (Colbo and Weller, 2007). This divergence of heat and salt is presumably achieved by the interaction of mesoscale and submesoscale processes with the surface layer, though the precise mechanisms are presently unclear. Candidates include oceanic eddies advecting relatively cold, fresh anomalies westward from the coastal zone and vertical mixing processes transporting heat and salt downward, across the base of the oceanic mixed layer. In general, however, little is know about the oceanic eddy processes in the SEP, not only regarding their role in influencing the mixed layer properties but also their potential role in modulating the concentration of aerosol precursors such as dimethylsulfide and complex organic species.
Clouds over the SEP exhibit a much stronger diurnal cycle of cloud cover and liquid water path, LWP (Rozendaal et al., 1995;Wood et al., 2002) than MBL clouds at comparable latitudes in the Northern Hemisphere. Regional model simulations (Garreaud and Muñoz, 2004) suggest that a large-scale diurnal subsidence wave formed by the interaction of the coastal jet along the Chilean coast with dry convective heating over the western Andean slopes travels at least 1000 km over the SEP and leads to a strong diurnal cycle of subsidence at remote locations. Using improved observations of how this wave influences the diurnal cycle of marine stratocumulus should be useful for assessing whether the diurnal variations of clouds in large scale models are well represented.

Motivation for the VOCALS regional experiment
The science issues described above are central to VOCALS (VAMOS Ocean-Cloud-Atmosphere-Land Study), an international CLIVAR program to develop and promote scientific activities leading to improved understanding, model simulations, and predictions of the southeastern Pacific (SEP) coupled ocean-atmosphere-land system, on diurnal to interannual timescales. VOCALS is ultimately driven by a need for improved numerical model simulations of the coupled climate system in both the SEP and over the wider tropics and subtropics. At the root of VOCALS's approach to the problem is the premise that its solution requires a synergy between numerical modeling, field studies, and extended observations such as buoys and satellites. With this in mind, the VOCALS Regional Experiment (VOCALS-REx) was conceived. In this manuscript we present an overview of the hypotheses, instrumentation, sampling platforms, sampling strategies, and missions conducted in pursuit of the science goals. We deliberately do not discuss any scientific results from REx; this paper is intended as a background framework and a supplement to the numerous other papers which present those results.

H1a
Variability in the physicochemical properties of aerosols has a measurable impact upon the formation of drizzle in stratocumulus clouds over the SEP.

H1b
Precipitation is a necessary condition for the formation and maintenance of pockets of open cells (POCs) within stratocumulus clouds.

H1c
The small effective radii measured from space in the coastal region of over the SEP are primarily controlled by anthropogenic, rather than natural, aerosol production, and entrainment of polluted air from the lower free-troposphere is an important source of cloud condensation nuclei (CCN).

H1d
Depletion of aerosols by coalescence scavenging is necessary for the maintenance of POCs.

H2a
Improvement of CGCMs performance in the SEP is key to the successful simulation of the ITCZ/SPCZ, complex, which will also benefit simulation of other regions. A significant improvement can be achieved through better representing the effects of stratocumulus clouds on the underlying surface fluxes and those of oceanic mesoscale eddies in the transport of heat.

H2b
Oceanic mesoscale eddies play a major role in the transport of relatively fresh water from the coastal upwelling region and in the production of sea-water and atmospheric DMS in the coastal and offshore regions. Upwelling, by changing the physical and chemical properties of the upper ocean, has a systematic and noticeable effect on aerosol precursor gases and the aerosol size distribution in the MBL over the SEP.

H2c
The diurnal subsidence wave ("upsidence wave") originating in northern Chile/southern Peru has an impact upon the diurnal cycle of clouds that is well-represented in numerical models.

H2d
The entrainment of relatively cool fresh intermediate water from below the surface layer during mixing associated with energetic near-inertial oscillations generated by transients in the magnitude of the trade winds is an important process to maintain heat and salt balance of the surface layer of the ocean in the SEP.
In Sect. 6, we will briefly summarize the VOCALS strategy for coordinating modeling work with REx (again, without presenting results). REx was designed to inform improvement of both global and regional climate and chemical transport models and also process-level models such as large-eddy simulations of aerosol-cloud interaction. In addition, REx made use of real-time output from several models for mission planning.
VOCALS-REx provided intensive observations of key processes contributing to the climate of the SEP. The observations are being used to help test a coordinated set of hypotheses presented in Table 1, to evaluate our ability to model the important physical and chemical processes in the SEP, and to help evaluate the performance of satellite retrievals. The VOCALS-REx hypotheses are organized into two broad themes: (1) the impacts of aerosols upon the microphysical and structural properties of stratocumulus clouds and drizzle production; (2) the coupled ocean-atmosphere-land system.

VOCALS-REx study region and dates
VOCALS-REx took place during October and November 2008, engaging over 150 scientists from 40 institutions in 8 nations. A variety of operations within a limited domain of the SEP coupled climate system were conducted (Fig. 2). REx operations took place in the domain 69-86 • W, 12-31 • S, with a concentration of sampling close to the 20 • S latitude line. This parallel was chosen as it transects the heart of the SEP stratocumulus sheet (Klein and Hartmann, 1993;George and Wood, 2010), exhibits strong longitudinal microphysical contrasts (Bennartz, 2007;Wood et al., 2008;George and Wood, 2010;Bretherton et al., 2010), crosses a region where open cell formation is frequently observed , and is impacted by mesoscale ocean eddies (e.g. Colbo and Weller, 2007;Toniazzo et al., 2009).
Overall, the VOCALS-REx period was characterized by near climatological atmospheric conditions off northern Chile and southern Peru . However, significant variations in MBL depth occurred during October when midlatitude troughs reached the VOCALS region leading to four episodes (1-2 day long) of mid-tropospheric upward motion. In contrast, November exhibited less synoptic forcing and almost continuous subsidence (Rahn and Garreaud, 2010b;Toniazzo et al., 2011). In the following sections we first discuss the research platforms and the instrumentation used to make observations during VOCALS-REx, followed by the chief mission types and sampling strategy.

Platforms and instrumentation
A total of five aircraft (NSF/NCAR C-130, the DoE G-1, the CIRPAS Twin Otter, the FAAM BAe-146, and the NERC Dornier 228, see Table 2) two research vessels (the NOAA R/V Ronald H. Brown, RHB, and the Peruvian IMARPE José Olaya, see Tables 3 and 4 respectively) sampled the lower atmosphere and upper-ocean during REx. These mobile platforms were complemented by a number of groundbased observational sites (Table 5).

Aircraft platforms
Three of the aircraft deployed in VOCALS-REx (C-130, G-1 and Twin Otter) were operational from 14 October to 15 November 2008, with the other two aircraft (BAe-146 and Do-228) operational from 26 October-15 November 2008. Table 2 shows the dates over which missions were flown, and Fig. 3 provides a graphical representation of the aircraft sampling as a function of day and longitude. Tables describing the specific aircraft missions are discussed below. The air-craft measurements are designed to critically address several of the VOCALS hypotheses (Table 1), particularly those related to aerosol-cloud-drizzle interactions and those involving the sources and sinks of atmospheric aerosols.

NSF/NCAR Lockheed C-130
The NSF/NCAR C-130Q is operated by the Research Aviation Facility (RAF) at the National Center for Atmospheric Research (NCAR) in the United States. During REx the C-130 flew missions up to 9 h in duration reaching 1600 km offshore, making it the longest range aircraft used in REx. The C-130 has a large payload and carries instruments and sensors in pods and pylons on both wings. Details of the instrumentation payload on the C-130 are given in Table 2. The aircraft is flown at an airspeed of approximately 100 m s −1 for boundary layer sampling. Details of the missions flown in REx are given in Table 6. Further information on the instrumentation on the C-130 including data quality can be found online at http://www.eol.ucar.edu/about/our-organization/raf/ data/vocals/vocals-documentation-summary.

FAAM BAe-146
The Facility for Airborne Atmospheric Measurements (FAAM) BAe-146 aircraft is operated by a joint agreement between the Met Office and the Natural Environment Research Council (NERC) in the United Kingdom. The BAe-146 served as the medium range aircraft operated in REx, flying missions of typically 5 hours and sampling up to 900 km offshore. The BAe-146 has a large payload and carries instruments and sensors in pods and pylons on both wings. Details of the instrumentation payload on the BAe-146 are given in Table 2. The aircraft is flown at an airspeed of approximately 100 m s −1 for boundary layer sampling. Details of the missions flown in REx are given in Table 7. Further information on the instrumentation on the BAe-146 including data quality can be found online at http: //data.cas.manchester.ac.uk/vocals/vocals-uk-summary.pdf.

DoE Gulfstream-1 (G-1)
The Department of Energy Gulfstream-1 (G-1) is operated by the Research Aircraft Facility (RAF) at the Pacific Northwest National Laboratory in the United States. The G-1 served as a medium range aircraft in REx, with sampling out to 800 km from the coast. The aircraft is flown at an airspeed of approximately 100 m s −1 for boundary layer sampling. Details of the instrumentation payload on the G-1 are given in Table 2. Details of the missions flown in REx are given in Table 8.
The NERC Dornier-228 is operated by the Airborne Research and Survey Facility (ARSF) of the Natural Environment Research Council (NERC) in the United Kingdom. Its main role in VOCALS-REx was remote sensing of clouds out Atmospheric state: temperature, humidity, winds/turbulence/turbulence dissipation, cloud and drizzle microphysics (FSSP at 200 Hz, CAPS probe in particle-by-particle mode) Remote sensing: UV zenith/nadir Aerosols: size distributions from 16-3000 nm (TSEMS, FIMS, PCASP), total CN (> 10 nm and > 2.5 nm), CCN (3 supersaturations, 0.18%, 0.26%, and 0.35%), aerosol composition (aerosol mass spectrometer, particle into liquid sampler, TRAC), scattering and absorption (3-wavelength nephelometer, 3-wavelength PSAP, Photo-thermal interferometer, Single particle soot photometer) to 76 • W, using lidar, a hyperspectral imager and polarimeter. Details of the instrumentation payload on the Do-228 are given in Table 2. Details of the missions flown in REx are given in Table 9. Most flights took place at an altitude of 4-5 km, with the remainder profiling the free troposphere to measure in-situ aerosol concentration. Typically, the Dornier Remote sensing: C-band radar reflectivity and Doppler winds within drizzle (3-d volumetric and range-height scans every 3 min, 60 km range); W-band radar reflectivity profiles and Doppler velocity for cloud/drizzle (vertically pointing 95 GHz cloud radar); cloud base and drizzle backscatter (lidar ceilometer); volumetric lidar backscatter and winds (scanning High Resolution Doppler Lidar, also operated in vertically pointing mode, 6 km range); liquid water path and water vapor path (23/31/90/183 GHz microwave radiometers); broad band irradiances.

MAX-DOAS)
Oceanography: 438 Underway CTD profiles (temperature, conductivity, pressure) to between 200 and 800 m depth, horizontal spacing from 1-30 km; 35 CTD profiles to 2500 m in and outside of eddies/fronts with associated water sampling for the collection of nutrients, salinity and oxygen samples; 10 SOLO profiling floats deployed with dissolved oxygen sensors; underway sea-surface salinity/temperature measurements; 19 surface drifters; 15 Vertical microstructure profiles (high resolution temperature, conductivity, velocity, pressure). overflew the flight path of the FAAM BAe146 with a similar airspeed (∼ 100 m s −1 ) and/or C-130 especially during the 20 • S missions (see below).

CIRPAS Twin Otter
The Twin Otter operated by the Center for Interdisciplinary Remotely Piloted Aircraft Studies (CIRPAS) was instrumented to make turbulence, cloud microphysics, and aerosol measurements (Table 2) in the near coastal region of the VO-CALS domain at 20 • S, 72 • W (a location termed here as Point Alpha, see Fig. 2). This relatively slow-moving aircraft (∼60 m s −1 ) made 5 h flights originating from Iquique Chile that allowed for 3 h of sampling at Point Alpha on 18 flights (Table 10).

Ship platforms
The two ships in VOCALS-REx sampled different locations at different times. The R/V Ronald H. Brown was operational for two phases, the first from 25 October to 2 November 2008 and the second from 10 November to 2 December 2008. The Peruvian R/V José Olaya operated from 2-17 October 2008. Figure 3 provides a graphical representation of the ship sampling as a function of day and longitude. Figures describing the specific ship sampling strategies are disussed below. The ship measurements are designed to critically address several of the VOCALS hypotheses (Table 1), particularly those related to the upper ocean, aerosolcloud-drizzle interactions, the physical and chemical interactions between the upper ocean and the lower atmosphere, and those involving the sources and sinks of atmospheric aerosols.

NOAA R/V Ronald H. Brown
The R/V Ronald H. Brown is operated by the National Oceanographic and Atmospheric Administration (NOAA), and served as the primary shipborne sampling platform for measurements in the vicinity of 20 • S from the coast out to 85 • W. The RHB also provided the means to deploy and recover moorings, drifters, and profiling floats during VO-CALS REx. The RHB payload was designed to sample both the upper ocean and the lower atmosphere during REx, and details are given in

Peruvian R/V José Olaya
The José Olaya is surveyed by the Instituto del Mar del Perú (IMARPE) and operated in Peruvian near-coastal waters to provide extensive sampling of the upper ocean, with additional atmospheric measurements (Table 4). The sampling Upper air: regular radiosonde launches predominately within an area about 200 km off the Pisco-San Juan region (see Fig. 8).
Oceanography: 113 CTD profiles (temperature, conductivity, pressure) to 1000 m depth in the coastal upwelling off southern Peru extending from the coast to 80-320 km, horizontal spacing from 19 km (nearshore) to 32-45 km (offshore). The CTD was deployed with dissolved oxygen and fluorescence sensors. Continuous records of VM-ADCP data (bin size 8 m, ping rate 0.3 s-1); underway sea surface temperature/salinity; 8 surface drifters. Collection of water samples for determination of oxygen, nutrients (phosphate, silicate, nitrate, nitrite), ph and chlorophyll-a concentrations in 78 stations. Underway measurements of partial pressure of carbon dioxide (pCO 2 ) complemented the biogeochemical observations.
Glider mission: continuous physical and biogeochemical data (temperature, salinity, dissolved oxygen, fluorescence and turbidity) were collected by a repeating section between 10 km and 100 km from the coast off Pisco. strategy (see below) was designed to examine the coastal upwelling region off Pisco-San Juan and extended from the Peruvian coast to 100-300 km offshore. The upper and lower atmosphere, the upper ocean property distribution and circulation, the biogeochemical characteristics, the plankton community structure as well as fishery responses were measured in a comprehensive, multidisciplinary basis. Details on the instrumentation onboard the Olaya are provided in Table 4. The National Center for Atmospheric Research (NCAR) Earth Observing Laboratory (EOL) deployed a GAUS (GPS Advanced Upper-air Sounding systems) radiosonde station on the José Olaya during VOCALS with sondes launched by IMARPE and IGP (Instituto Geofísico del Perú) and IRD (Institut pour le Récherche et Devéloppement) scientists. A total of 133 soundings were launched at varying intervals from 30 September to 17 October 2008. The launch sites were predominately within an area about 200 km off the coast of the Ica region of southern Peru. Vaisala RS92G radiosondes were used throughout.

Paposo
Extensive aerosol and meteorological measurements were made at two sites near Paposo (25 • 01 S, 70 • 28 W) on the Northern Chilean coast (see map Figs. 2 and 4). In terms of the flow in the MBL, Paposo sits upwind of the primary focus area along the 20 • S parallel and the measurements are designed to help constrain the physical and aerosol properties of airmasses leaving the continent to be advected over the broader SE Pacific region. Two sites were used near Paposo (Table 5)  The elevated Paposo site is close to the peak of the hill in the coastal range in which it is situated (Fig. 4). , the upper site was almost continually within the marine boundary layer (MBL), although earlier in the season the inversion was occasionally lower which allowed sampling above the MBL. Table 5 details the measurements made at the upper Paposo site. Aerosol sampling was carried out using a custom-made multidirectional aerosol inlet and a multiport sampling configuration (see Fig. 4), with additional sampling lines for aerosols during 4-15 November. The primary sampling line was used to connect with the scanning mobility particle spectrometer (SMPS), optical particle counter (OPC), nephelometers, aethalometer and ozone analyzer, and the same line was used to sample submicron (< 1 µm diameter) aerosols on filters for chemical analysis. Aerosol filter measurements are described further in Chand et al. (2010). Meteorological and radiation measurements at the upper site were made by the University of Chile, and these measurements are described further in (Muñoz et al., 2011). At the lower Paposo site, an eye-safe 1.574 µm lidar, a weather station, and a sounding system were installed at the Paposo foothill site near the coast (Table 5 and map Fig. 4). The lidar was primarily vertically-pointing but some slant path scans were also performed. An identical set of meteorological parameters to that measured at the upper site was measured at the lower site. Multiple soundings per day were made from the site (Table 5 provides details of the launch times).

Paranal
A suite of aerosol measurements (see

IMET Buoy
The Improved Meteorology (IMET) moored buoy is situated at approximately 20 • S, 85 • W (see Table 5 for precise location) at the western end of the sampling conducted during VOCALS-REx. The mooring has been operational since October 2000 and has provided an excellent intermediate-term record of both the surface meteorology/radiation, and the upper ocean thermodynamic and dynamic structure. The meteorological and radiation measurements (Table 5) on the IMET buoy are described and their performance evaluated in Colbo and Weller (2009). The upper ocean measurements include temperature profiles, sea-surface temperature, salinity and currents. Further details can be found on the WHOI Upper Ocean Processes website (Table 5).

DART/SHOA Buoy
The Deep-ocean Assessment and Reporting of Tsunamis/Servicio Hidrográfico y Oceanográfico de la Armada de Chile (DART/SHOA) moored buoy at approximately 19.5 • S, 74 • W (see Table 5 for precise location) has been instrumented with meteorological and oceanographic measurements from October 2006 through January 2010.
Meteorological measurements similar to those on the IMET buoy (Colbo and Weller, 2009) were made during much of this period. Upper ocean measurements of temperature and salinity at 14 depths were also made from 2006 onwards.

Matching sampling strategy to the VOCALS hypotheses
The REx sampling strategy was carefully designed and coordinated between platforms to test key VOCALS hypotheses listed in Table 1. Approximately half way through the field phase, when the RHB was in between its two cruise legs, an "all-hands" meeting was held in Arica to discuss progress and strategize about the needs for sampling during the remainder of the campaign. Some important adjustments to the sampling strategy were made at this point. Briefly, the VOCALS aerosol-cloud-drizzle hypotheses can be paraphrased as -H1a: aerosol variations significantly affect drizzle formation.
-H1b/d: drizzle-induced aerosol scavenging is required for POC formation and maintenance.
-H1c: cloud droplet radii are smaller near the coast due to anthropogenic aerosol emissions from South America.
In the REx region, both repeated surveying and sampling of specific features are useful for testing these hypotheses. Repeated sampling of the persistent gradients in aerosols, clouds and precipitation between nearshore and offshore regimes allows robust features of the gradient region to stand out and can be used to study correlations among aerosols, cloud macrostructure, and meteorology present on individual days. POCs present extreme examples (typically POCs are among the very cleanest and most strongly drizzling of airmasses in the SEP) that challenge our physical understanding of cloud-aerosol-precipitation interaction. Thus, an aircraft sampling strategy mainly focused on repeated sampling across the aerosol gradient region (with a few missions at the end parallel to the coast to characterize the offshore aerosol distribution upwind of the main VOCALS-REx study region), interspersed by opportunistic sampling of any POCs  within range. Because of the desire for repeated sampling strategies, the aircraft favored particular times of day and did not attempt to characterize the diurnal variability of the cloud-topped boundary layer. The ship, mooring, and landbased sampling was aimed at complementing the aircraft through a better characterization of diurnal variability along 20 • S (particularly with the RHB and the Iquique sounding site) and of the upstream anthropogenic aerosol sources (the Paposo and Paranal land sites).
The observational VOCALS coupled ocean-atmosphere hypotheses in Table 1 can be summarized: -H2b/d: the offshore ocean mixed layer SST and salinity are decreased by ocean mesoscale eddy transports and entrainment from below. Oceanic DMS affects the boundary layer aerosol both in the upwelling zone and far offshore.
-H2c: a subsidence wave driven by slope heating on the Andes measurably affects the diurnal cycle of stratocumulus.
VOCALS-REx tackled these hypotheses mainly with a shipbased strategy (RHB and José Olaya), through sampling of clouds and atmospheric profiles through the diurnal cycle for H2c and survey-style sampling for H2b/d.   For both sets of hypotheses, IMET/DART mooring observations, satellite observations, and modeling on a range of scales are envisioned as vital complements to the in-situ observations.

Aircraft missions
The following aircraft mission strategies were used during VOCALS-REx:  Fig. 6. Cross Section missions summary as a function of date and longitude along 20 • S. Color coding shows the approximate local/UTC time of sampling. Times for which Cross-Section mission data is available are provided at right. Individual aircraft flight numbers are also given. Missions with missing outbound or return legs indicate that the aircraft was involved in a different mission for part of its flight.
1. Cross-Section (XS) missions along 20 • S latitude (or other proximal latitudes) from the coast to close to the IMET buoy at 85 • W (mission plan shown in Fig. 5) aimed to sample longitudinal gradients in clouds, the MBL, and aerosols. A total of 12 Cross-Section missions were flown along 20 • S during REx (mission details shown in Fig. 6), with more flown along nearby latitudes (especially by the G-1 aircraft, see Table 8). Emphasis in these missions was on good sampling within the MBL and the air in the lowermost part of the free troposphere that would be entrained into the MBL, but profile measurements were also made up to 3-4 km to capture aerosol layers and the vertical thermodynamic structure aloft. The BAe-146 missions also  Fig. 7, and a summary of the POC sampling from the various platforms is given in  3. Stacked cloud and/or radiation missions in which one or two aircraft sample a cloudy boundary layer airmass, typically using stacked legs 50-100 km in length. For two-aircraft missions, the upper aircraft primarily served as a radiation/remote sensing platform and flew in the free troposphere. All the aircraft other than the C-130 carried out missions of this type. All Twin Otter missions were of this type, and additionally were carried out at the same location (at so-called "Point Alpha", 20 • S, 72 • W); 4. Pollution Survey missions in which aircraft sampled within a few hundred km of the Peruvian and Chilean coasts, with the aim of characterizing the lower atmosphere in the vicinity of pollution source regions. These missions replaced the polluted Lagrangian missions which had been originally planned for VOCALS-REx, because it became clear that Lagrangian missions starting around 20-25 • S would not sample sufficiently close to the major pollution sources to capture the aging process. Emphasis was placed on sampling both within the MBL and in the lower free troposphere, but occasional profiles up to 3-4 km altitude were also employed. Figure 2 shows the typical locations of these flights. The BAe-146 and C-130 aircraft performed the bulk of the pollution survey sampling, with six flights dedicated to this mission type; 5. Intercomparison flights, either aircraft-aircraft (at several different flight levels both in and above the MBL) or to compare aircraft and ship measurements (with the aircraft flying at its lowest possible flight level, typically 150 m). The summary of intercomparisons is given in Table 12. Tables 6, 7, 8, 9, and 10 present the specific missions flown by the C-130, BAe-146, G-1, Do228, and Twin Otter respectively. Figure 8 shows the track of the R/V José Olaya during the VOCALS REx cruise (2-17 October 2008). A total of 133 radiosonde soundings were acquired at varying spatiotemporal intervals from 30 September to 17 October 2008. Launch sites were predominately within the upwelling zone, about 200 km from the coast of the Pisco-San Juan upwelling region. Temperature, salinity and currents were measured to characterize the physical properties of the upwelling plume and the associated thermal front. A cluster of 8 surface drifters were deployed across the upwelling front in order to study the advective and diffusive processes inside this feature. The glider (autonomous underwater vehicle) mission was designed to examine the high-resolution structure and dynamics of the upwelling plume and thermal front off Pisco between 10 km and 100 km from the coast. The distribution of biogeochemical and biological parameters as well as fish abundance were also sampled to study the feedback of ocean/atmosphere interactions on biological and fishery activity.  Legs 1 and 2 involved studies of the ocean, the atmosphere, and their coupling as part of VOCALS-REx. Leg 1 focused primarily upon measurements at the IMET and SHOA buoys, while Leg 2 involved more surveying of mesoscale ocean features. The sampling strategy for Leg 2 was optimized during the "all-hands" meeting in Arica between legs, the adjustments being necessary because operational delays reduced the Leg 1 sampling by almost 10 days which compromised the mesoscale surveying during Leg 1. Coordinated sampling with research aircraft working from Arica and Iquique took place during both cruises, with the majority of coordinated sampling taking place during Leg 2 (see Tables 6 to 9 for details of RHB-aircraft cosampling. A total of 210 radiosondes were obtained at 4 h intervals within the marine stratocumulus region. The ship sampled multiple times across relatively sharp transitions of cloud coverage including clear to broken to overcast stratocumulus cloud conditions. It was overcast approximately 80% of the time. Drizzle was prevalent: drizzle-containing cells with significant radar reflectivity (Z > 0 dBZ) were observed within a 60 km radius of the ship roughly half the time. The RHB research cruise for VOCALS-Rex was designed to address important aspects of both (1) aerosol-cloud-drizzle hypotheses and (2) coupled ocean-atmosphere hypotheses.

NOAA R/V Ronald H. Brown
Aerosol-cloud-drizzle interactions vary in both space and time at a multitude of scales. The ship provided a platform to investigate in detail aerosol distributions and composition, including diurnal patterns during slow transects over much smaller regions of the marine boundary layer than is covered by the aircraft in a single hour. For this reason, the nearly 60-day cruise provided measurements of marine aerosol with a greater range of statistical variability, more chemical detail (such as organic functional groups, see Hawkins et al., 2010), and highly accurate standards as references (such as ion chromatography) that are complementary to the aircraftbased data sets. In addition, the measurement of radon on the RHB permits an assessment of the time that airmasses have spent over the ocean. The RHB studies of aerosol properties provide a comprehensive basis for addressing the variability in physicochemical properties. These measurements also serve as the basis for comparison of the sources and composition of the aerosol particles, providing comprehensive information with which to compare to satellite-retrieved properties.
The main objectives of the oceanographic field work conducted from the R/V Ron Brown (Legs 1 and 2) were: (i) to map the mean and eddy (mesoscale/submesoscale) temperature, salinity and velocity distribution within the SEP's upper ocean during VOCALS-REx; (ii) to deploy Lagrangian floats and drifters within the SEP; (iii) to recover and re-deploy the STRATUS and DART moorings. The synoptic survey across the SEP region included the collection of 35 CTD (Conductivity, Temperature, Depth profiles) up to 2000 m depth, and of 438 UCTD (Underway CTD) profiles, ranging between 200 and 800 m deep, to map the meridional distribution of properties across the SEP along three distinct latitude lines (Fig. 9). During the surveys, spatial and temporal sampling was increased to resolve a number of oceanic fronts and eddies, including 4 cyclones, 2 anticyclones, the coastal currents and upwelling front at 21.5 • S. Microstructure profiles to quantify mixing rates were obtained using a Vertical Microstructure Profiler (VMP) at 15  10 profiling Lagrangian floats (SOLO floats, equipped with an oxygen sensor) and 19 surface drifters -some in eddies and some throughout the SEP -were designed to provide long-term context to the synoptic measurements -together with the instruments recovered and re-deployed on the IMET/STRATUS and DART buoys.

Visible Infrared Solar-Infrared Split Window Technique (VISST)
Cloud and radiation parameters at 4-km resolution were derived from the tenth Geostationary Operational Environmental Satellite imager (GOES-10), located at 60 • W, using techniques developed at NASA Langley Research Center (LaRC). The GOES-10 data were analyzed every half hour for the region bounded by 10 • S, 30 • S, 65 • W, and 90 • W for the period between 11 September and 1 December 2008 and provided in near-real time for mission planning and analysis. Clouds were detected using the method of Minnis et al. (2008) and cloud properties were retrieved during the daytime using the Visible Infrared Shortwave-infrared Split-window Technique (VISST). At night, cloud properties were retrieved using the Shortwave-infrared Infrared Splitwindow Technique (SIST). The methods are described in detail for application to MODerate-resolution Imaging Spectroradiometer (MODIS) data by Minnis et al. (2008Minnis et al. ( , 2010. The VISST uses the 0.65, 3.9, 10.8, and 12.0 µm channels, while the SIST uses the same channels minus the 0.65 µm data. The GOES-10 0.65 µm channel was calibrated against the Terra MODIS 0.64 µm channel using the technique of Minnis et al. (2002). The available derived parameters and means of accessing the data are similar to those described by (Palikonda et al., 2006). Both pixel-level and 0.5 × 0.5 • averages are available each hour in image and digital form 2 . The VISST and SIST assume that only single-layer clouds are in a given pixel. In addition to the standard approach described by Minnis et al. (2010), cloud-top height and pressure were also retrieved using the method described by Zuidema et al. (2009). One additional parameter, a multilayer cloud identifier was computed for each pixel using the approach of Pavolonis and Heidinger (2004). In addition to the cloud properties, spectral radiances and estimates of the top-of-atmosphere shortwave albedo and outgoing longwave radiation are included. Figure 10 shows an example of three parameters for a GOES-10 image taken at 15:45 UTC, 27 October 2010. The pseudo-color RGB image (Fig. 10a) shows low clouds in the orange and peach shades with high cirrus clouds appearing white, gray, and magenta. The effective cloud temperatures T c are displayed in Fig. 10b for an abbreviated range of 273 K < T c <300 K to better show variations in stratocumulus cloud temperatures. Temperatures less than 273 K are indicated in the maroon shade. For this case, T c ranges from 274 K to 284 K for the marine stratocumulus clouds. Smaller values are evident where thin cirrus clouds occur over the low clouds. Cloud optical depths (Fig. 10c) range from less than 1 at some cloud edges to more than 40 near 18 • S, 78 • W. The VISST-derived droplet effective radius, re, (Fig. 10d) varies from about 7 to 25 µm across the scene with most of the largest values occurring around the edges of the POCs. The smallest droplets are mostly near the coast. The pixel-level products, exemplified in Fig. 10a-d, are used to produce 0.5 × 0.5 • regional means at each half hour for many of the cloud and radiation parameters (Palikonda et al., 2006). Examples of 0.5 × 0.5 • regionally-averaged cloud top-height Z t and liquid water path (LWP) are shown in Fig. 10e, f. The Z t values estimated as in Minnis et al. (2010) range from less than 1 km up to more than 3 km in the southwestern portion of the domain. Higher clouds near the center of the domain correspond to the thin cirrus clouds over the stratocumulus deck. The heights based on the Zuidema et al. (2009) technique are generally lower (not shown). The cloud LWP ranges from less than 50 g m −2 along the coast to over 200 g m −2 near the center of the domain. The LaRC cloud properties are based on near-real time retrievals. A refined dataset using the latest GOES-10 calibrations, a higher resolution sea surface temperature dataset, and algorithm updates is being generated to provide a more accurate set of cloud properties for stratocumulus research and for comparison with the other experiment measurements to better define the uncertainties in the satellite products.

Gridded cloud cover product from the University of Manchester/Met Office
Thermal infrared data from GOES-10 (Channel 4, 10.7 µm), converted to netCDF format and archived on the VOCALS data archive (see Sect. 8), have been used to generate a dataset documenting the variability in cloud amount during the VOCALS-REx period. The GOES-10 data were analyzed between 1 October and 8 December 2008 in a region from 3.5-31.5 • S and 68.5-96.5 • W. Note that this is a more extensive region than for the VISST GOES-10 products described above. Clouds are classified on all available GOES-10 scans (typically every 15 to 30 min) at a horizontal resolution of 4 km, and cloud cover fractions are gridded at 0.25 × 0.25 • resolution. Further details are given in Abel et al. (2010), and the dataset is available on the VOCALS archive, described below.

MODIS subset
A dedicated subset of MODIS imagery from NASA's Terra and Aqua satellites for the VOCALS-REx study region is available for browsing on the MODIS Rapidfire website http://rapidfire.sci.gsfc.nasa.gov/subsets/?subset=VOCALS.

Overview of the VOCALS modeling program
An overarching goal of VOCALS is to improve model simulations of key climate processes using the SEP as a testbed, particularly in coupled models that are used for climate change projection and ENSO forecasting. Hence, REx was developed in close coordination with the VOCALS modeling program, whose main goals are: 1. Understanding and reducing the warm SEP SST bias near the coast and excessive interhemispheric symmetry in the eastern tropical Pacific present in most coupled climate models.
2. Using the SEP as a testbed for better simulation of boundary layer cloud processes and aerosol-cloud interaction, including the relative roles of natural and anthropogenic aerosol sources and their impact on cloud optical properties (coverage, thickness, and droplet size).
3. Improving the understanding and simulation of oceanic budgets of heat, salinity, and nutrients in the SEP and their feedbacks on the regional climate.
4. Elucidating interactions between the SEP and other parts of Earth's climate system, including the South American continent, the Pacific circulation and ENSO.
The VOCALS modeling vision is based on the concept of a multiscale hierarchy of models, both in time and space. This is motivated by the multiscale nature of processes in the SEP and the multiscale hierarchy of VOCALS observations, including REx, extended in-situ and satellite data. In this spirit, one VOCALS modeling goal is to test models used for long-term climate projection as rigorously as possible by applying them on a different timescale, namely the short period of intensive data gathering during REx, by testing them in a weather forecasting mode. Another goal is to compare observations with models of various horizontal and vertical resolutions, e.g. higher-resolution regional models vs. coarserresolution global models. A third goal is to test and apply small-scale process models, e.g. large-eddy simulation (LES) models of the cloud-topped boundary layer, which can inform our physical understanding and help guide the development of parameterizations for larger-scale models.
Many aspects of REx were designed to facilitate these modeling goals. The atmospheric observation strategy included repeated airborne and ship-based measurements along one transect, 20 • S, to facilitate comparison with global and regional atmospheric models used in forecast mode and to provide a rough climatology that could be compared with a broader group of coupled ocean-atmosphere models. All aircraft included cloud physics, turbulence and aerosol/chemical composition measurements for testing the representation of aerosol/cloud interaction in models; the Wyoming Cloud Radar on the C-130 also added precipitation profiles into this dataset. This integrated suite of measurements provides a strong constraint on simulations of SEP clouds and aerosols. Synthesis papers by Bretherton et al. (2010) (boundary layer and physical cloud properties) and Allen et al. (2011) (aerosol and chemical composition) summarize the results of the REx 20 • S measurements in multiplatform 20 • S synthesis datasets that will be part of the EOL VOCALS data archive and are designed to be convenient for comparison with large-scale models. Modeling studies by Rahn and Garreaud (2010a,b) and Abel et al. (2010) focus on comparison with REx 20 • S measurements. The VO-CALS assessment (see Sect. 6.3) was also conceived as an integrated part of REx, and will make use of the REx 20 • S synthesis datasets.
The REx POC missions were designed for comparison with large-domain LES of cloud-aerosol-precipitation interaction, and modeling papers utilizing these REx datasets are already emerging, e.g. Wang et al. (2010).
REx ship-based sampling of mesoscale ocean eddies was also envisioned to complement and test a regional eddyresolving (5 km resolution) ocean model run in data assimilation mode; such modeling efforts are underway.

Real-time modeling during REx
Several modeling groups supported VOCALS-REx mission planning and field data interpretation through provision of plots from real-time forecasts. This also provided those groups a good opportunity to evaluate their models in the field. These forecast products are archived in the VOCALS-REx Field Data Catalog 3 ; they remain a useful "quick-look" resource. They include plots of simulated regional meteorological fields, vertical profiles, trajectories and cross-sections of selected fields and chemical constituents, especially along 20 • S, and some regional zonal cross-sections through the upper ocean.
A synopsis of contributed and archived real-time products follows.

UKMO
The UK Met Office (UKMO) submitted 20 • S cross-sections and horizontal maps of a variety of fields from their operational weather forecast model, the Unified Model (UM), run at 40 km resolution globally, and from a 17 km resolution regional version of the UM nested inside their global model. In addition they contributed real-time forecasts from their Numerical Atmospheric Dispersion Modeling Environment

VOCALS data management
The NCAR/EOL provided data management support, coordination, and a long-term archive for VOCALS datasets. Details regarding VOCALS Data Management can be found on the VOCALS Project web page 7 . This web page contains the VOCALS data policy, instructions for data submission, relevant documentation, links to related projects data, and access to the distributed VOCALS long-term archive [i.e. Master List (ML) of VOCALS International Datasets]. The ML contains direct access to all datasets organized by data category and data source site with associated dataset documentation. In addition, the VOCALS-Rex Field Catalog 8 used during the field phase to provide operations and mission/scientific reports, operational and preliminary research imagery/products is available as a browse tool for use by researchers in the post-field analysis phase and is included as part of the archive.

Conclusions
The VOCALS Regional Experiment (VOCALS-REx) was an international field experiment designed to examine critical aspects of the coupled climate system of the Southeast Pacific region. VOCALS-REx took place during October and November 2008 in a domain 69-86 • W, 12-31 • S. Sampling with a variety of platforms including two ship, five research aircraft, land sites and two instrument moorings will ensure that researchers have a number of different observational angles with which to test the VOCALS hypotheses. The purpose of this paper is to bring together in one document the scientific goals, the platforms and instrumentation, and the sampling strategies employed during the program. It is hoped that this will serve the VOCALS research community by providing a central location that describes the essence of the field program. Perhaps more importantly, we hope that it will help to provide an important legacy that will be available to researchers over the coming years.