• Alizadeh-Choobari, O., P. Zawar-Reza, and A. Sturman, 2014: The “wind of 120 days” and dust storm activity over the Sistan Basin. Atmos. Res., 143, 328341, https://doi.org/10.1016/j.atmosres.2014.02.001.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Arctic Submarine Laboratory, 2018: ICEX 2018 briefing book. Arctic Submarine Laboratory Doc., 29 pp.

  • Barlow, M., B. Zaitchik, S. Paz, E. Black, J. Evans, and A. Hoell, 2016: A review of drought in the Middle East and southwest Asia. J. Climate, 29, 85478574, https://doi.org/10.1175/JCLI-D-13-00692.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bretherton, F. P., R. E. Davis, and C. B. Fandry, 1976: A technique for objective analysis and design of oceanographic experiments applied to MODE-73. Deep-Sea Res. Oceanogr. Abstr., 23, 559582, https://doi.org/10.1016/0011-7471(76)90001-2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chu, J. H., A. Levine, S. Daida, D. Schiber, E. Fukuda, and C. R. Sampson, 2017: The Joint Typhoon Warning Center tropical cyclone best-tracks, 1945–2017. Naval Meteorology and Oceanography Command, accessed 1 March 2020, https://www.metoc.navy.mil/jtwc/jtwc.html?best-tracks.

  • Coral Reef Research Foundation, 2014: A summary of Palau’s typhoon history 1945-2013. Coral Reef Research Foundation Tech. Rep., 17 pp.

  • Daley, R., 1991: Atmospheric Data Analysis. Vol. 2. Cambridge University Press, 471 pp.

  • Davis, R. E., 1985: Objective mapping by least squares fitting. J. Geophys. Res., 90, 47734777, https://doi.org/10.1029/JC090iC03p04773.

  • Duffie, W., 2015: ONR-sponsored technology aids recovery of Alaska plane wreck. ONR, https://www.onr.navy.mil/en/Media-Center/Press-Releases/2015/XMET-Used-In-Alaska-Plane-Crash-Recovery.

  • Emery, W. J., and R. J. Thomson, 2001: Data Analysis Methods in Physical Oceanography. Gulf Professional Publishing, 638 pp.

  • Esmaili, O., and M. Tajrishy, 2006: Results of the 50 year ground-based measurements in comparison with satellite remote sensing of two prominent dust emission sources located in Iran. Proc. SPIE, 6362, 636209, https://doi.org/10.1117/12.692989.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fiebrich, C. A., 2009: History of surface weather observations in the United States. Earth-Sci. Rev., 93, 7784, https://doi.org/10.1016/j.earscirev.2009.01.001.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fuller, J. F., 1974: Weather and war. U.S. Air Force Air Weather Service Rep., 25 pp.

  • Fuller, J. F., 1990: Thor’s Legions. Amer. Meteor. Soc., 443 pp.

  • Goudie, A. S., and N. J. Middleton, 2006: Desert Dust in the Global System. Springer, 287 pp.

  • Gouezo, M., Y. Golbuu, R. van Woesik, L. Rehm, S. Koshiba, and C. Doropoulos, 2015: Impact of two sequential super typhoons on coral reef communities in Palau. Mar. Ecol. Prog. Ser., 540, 7385, https://doi.org/10.3354/meps11518.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hamidi, M., M. R. Kavianpour, and Y. Shao, 2013: Synoptic analysis of dust storms in the Middle East. Asia-Pac. J. Atmos. Sci., 49, 279286, https://doi.org/10.1007/s13143-013-0027-9.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Holthuijsen, L. H., N. Booji, and L. Bertotti, 1996: The propagation of wind errors through ocean wave hindcasts. J. Offshore Mech. Arct. Eng., 118, 184189, https://doi.org/10.1115/1.2828832.

    • Search Google Scholar
    • Export Citation
  • Joint Staff, 2018: Meteorological and oceanographic operations. U.S. Joint Chiefs of Staff Publ. 3-59, 90 pp.

  • Kaskaoutis, D., A. Rashki, E. Houssos, A. Mofidi, D. Goto, A. Bartzokas, P. Francois, and M. Legrand, 2014: Meteorological aspects associated with dust storms in the Sistan region, southeastern Iran. Climate Dyn., 45, 407424, https://doi.org/10.1007/s00382-014-2208-3.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kayanne, H., H. Yamano, and R. H. Randall, 2002: Holocene sea-level changes and barrier reef formation on an ocean island, Palau Islands, western Pacific. Sediment. Geol., 150, 4760, https://doi.org/10.1016/S0037-0738(01)00267-6.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Liu, C. C., T. Y. Shyu, T. H. Lin, and C. Y. Liu, 2015: Satellite-derived normalized difference convection index for typhoon observations. J. Appl. Remote Sens., 9, 096074, https://doi.org/10.1117/1.JRS.9.096074.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lyapustin, A., Y. Wang, I. Laszlo, R. Kahn, S. Korkin, L. Remer, and J. S. Reid, 2011: Multiangle Implementation of Atmospheric Correction (MAIAC): 2. Aerosol algorithm. J. Geophys. Res., 116, D03211, https://doi.org/10.1029/2010JD014986.

    • Search Google Scholar
    • Export Citation
  • Middleton, N. J., 1986: Dust storms in the Middle East. J. Arid Environ., 10, 8396, https://doi.org/10.1016/S0140-1963(18)31249-7.

  • Miller, S. D., 2003: A consolidated technique for enhancing desert dust storms with MODIS. Geophys. Res. Lett., 30, 2071, https://doi.org/10.1029/2003GL018279.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Miller, S. D., A. P. Kuciauskas, M. Liu, Q. Ji, J. S. Reid, D. W. Breed, A. L. Walker, and A. A. Mandoos, 2008: Haboob dust storms of the southern Arabian Peninsula. J. Geophys. Res., 113, D01202, https://doi.org/10.1029/2007JD008550.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Nelson, M. S., 2016: Short-range acoustic propagation under Arctic ice cover ICEX-16. M.S. thesis, Dept. of Oceanography, Naval Postgraduate School, 55 pp.

  • OFCM, 2017: Surface weather observations and reports. OFCM Doc., 98 pp.

  • Ponce de León, S., and F. J. Ocampo-Torres, 1998: Sensitivity of a wave model to wind variability. J. Geophys. Res., 103, 31793201, https://doi.org/10.1029/97JC02328.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Prasad, A. K., and R. P. Singh, 2007: Changes in aerosol parameters during major dust storm events (2001–2005) over Indo-Gangetic plains using AERONET and MODIS data. J. Geophys. Res., 112, D09208, https://doi.org/10.1029/2006JD007778.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ranjbar, M., and F. Iranmanesh, 2008: Effects of “drought” on “wind eroding and erosion” in Sistan region with use of satellite multiple images. Islamic Azad Univ., Shahre Rey Branch, 14–30.

  • Rashki, A., and D. Kaskaoutis, C. J. de W. Rautenbach, and P., Eriksson, 2012: Changes of permanent lake surfaces, and their consequences for dust aerosols and air quality: The Hamoun Lakes of the Sistan area, Iran. Atmospheric Aerosols—Regional Characteristics—Chemistry and Physics, H. Addul-Razzak, Ed., InTech, 162–202, https://doi.org/10.5772/48776.

    • Crossref
    • Export Citation
  • Reynolds, R. M., 2002: Meteorology and climate. The Gulf Ecosystem: Health and Sustainability. N. Y. Khan, M. Munawar, and A. R. G. Price, Eds., Backhuys Publishers, 41–51.

    • Crossref
    • Export Citation
  • Ridley, D., C. Heald, J. Pierce, and M. J. Evans, 2013: Toward resolution-independent dust emissions in global models: Impacts on the seasonal and spatial distribution of dust. Geophys. Res. Lett., 40, 28732877, https://doi.org/10.1002/grl.50409.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rogowski, P., J. Zavala-Garay, K. Shearman, E. Terrill, J. Wilkin, and T. H. Lam, 2019: Air-sea-land forcing in the Gulf of Tonkin: Assessing seasonal variability using modern tools. Oceanography, 32 (2), 150161, https://doi.org/10.5670/oceanog.2019.223.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Saeed, T. M., and H. Al-Dashti, 2011: Optical and physical characterization of “Iraqi Freedom” dust storm, a case study. Theor. Appl. Climatol., 104, 123137, https://doi.org/10.1007/s00704-010-0334-3.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shao, Y., 2008: Physics and Modelling of Wind Erosion. Springer, 452 pp.

  • Teixeira, J. C., M. P. Abreu, and C. Guedes Soares, 1995: Uncertainty of ocean wave hindcasts due to wind modeling. J. Offshore Mech. Arctic Eng., 117, 294297, https://doi.org/10.1115/1.2827237.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tolman, H. L., 1999: User manual and system documentation of Wavewatch-III version 1.18. NOAA/NCEP Tech. Note 166, 110 pp.

  • Vishkaee, F. A., C. Flamant, J. Cuesta, P. Flamant, and H. R. Khalesifard, 2011: Multiplatform observations of dust vertical distribution during transport over northwest Iran in the summertime. J. Geophys. Res., 116, D05206, https://doi.org/10.1029/2010JD014573.

    • Search Google Scholar
    • Export Citation
  • Vishkaee, F. A., C. Flamant, J. Cuesta, L. Oolman, P. Flamant, and H. R. Khalesifard, 2012: Dust transport over Iraq and northwest Iran associated with winter shamal: A case study. J. Geophys. Res., 117, D03201, https://doi.org/10.1029/2011JD016339.

    • Search Google Scholar
    • Export Citation
  • Wilderson, W. D, 1991: Dust and sand forecasting in Iraq and adjoining countries. USAF Environmental Technical Applications Center Rep., 72 pp.

  • Wood, L. E., 1946: Automatic weather stations. J. Meteor., 3, 115121, https://doi.org/10.1175/1520-0469(1946)003<0115:AWS>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery
    Fig. 1.

    A massive dust storm (haboob) envelops the military camp at Al Asad, Iraq, on 27 Apr 2005. Department of Defense (DoD) photo by Corporal Alicia M. Garcia, U.S. Marine Corps (released).

  • View in gallery
    Fig. 2.

    (a) Image of modular XMET components packed in a ruggedized shipping case. (b) Pictorial illustrations showing assembly instructions of XMET components. (c) Computer-aided design (CAD) image of assembled XMET system with optional ceilometer.

  • View in gallery
    Fig. 3.

    XMET systems communicate through Iridium’s bidirectional SBD protocol and a local mail server. Data are processed immediately upon delivery and made available through a variety of services in order to accommodate several endpoints.

  • View in gallery
    Fig. 4.

    Time series of visibility (orange) and wind speed (gray) observations from an XMET deployed at the Al Asad Air Base in July 2008.

  • View in gallery
    Fig. 5.

    Time series of (a) wind speed, (b) temperature, (c) barometric pressure, and (d) system voltage observations from an XMET deployed in western Afghanistan from 2011 February through October 2017 (2443 days). A total of 58 145 weather observations were collected with reporting success of 99.2%.

  • View in gallery
    Fig. 6.

    Global map showing XMET deployment locations since 2008. Deployment locations discussed in this paper are labeled.

  • View in gallery
    Fig. 7.

    XMETs deployed (a) in desert conditions, (b) in a maritime environment, (c) on a glacier, and (d) on sea ice.

  • View in gallery
    Fig. 8.

    (a) Map of the primary islands in the Palau archipelago and the deployment locations of two XMET systems. (b) Map of the Vietnamese coastline with locations of three environmental installations denoted by the black triangles. An XMET system was integrated into each environmental site (see inset).

  • View in gallery
    Fig. 9.

    Topographic map of southwestern Afghanistan with XMET deployment locations and wind observations marked by black squares and overlaid with 1-km resolution MODIS AOD products for (a) 31 and (c) 27 Jul 2010 and (b),(d) their corresponding objectively mapped visibility and wind estimates. Objectively mapped estimates that have ratios of error to signal variance smaller than 0.8 are plotted. The primary northwesterly wind direction (thick black arrow) and dust source regions (red dotted line) are denoted in (a). Two southern XMET sites in (a) are labeled (XM1, XM2) for discussion in the manuscript. Visibility contours derived from XMET observations are included for reference between MODIS and OA plots (black lines).

  • View in gallery
    Fig. 10.

    Map of the western Pacific overlaid with the trajectories of Typhoons (a) Bopha and (b) Haiyan. The dotted black box denotes a region illustrated within the inset showing the wind circulation pattern of each storm as it approached two XMET locations (XM20 and XM44). Wind quiver arrows are color coded by intensity. Black quiver arrows at XM20 and XM44 denote speed and direction of observed XMET winds. Typhoon trajectories and maximum sustained wind speeds are provided by the JTWC unless otherwise noted, and were downloaded from https://www.metoc.navy.mil/jtwc/jtwc.html?best-tracks/.

  • View in gallery
    Fig. 11.

    Palau XMET time series of observed (a),(d) maximum wind speed, (b),(e) direction, and (c),(f) barometric pressure during the passage of Typhoons (a)–(c) Bopha and (d)–(f) Haiyan.

All Time Past Year Past 30 Days
Abstract Views 382 26 0
Full Text Views 355 261 23
PDF Downloads 314 222 14

XMET—An Unattended Meteorological Sensing System for Austere Environments

Peter Rogowski Coastal Observing Research and Development Center, Marine Physical Laboratory, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, California

Search for other papers by Peter Rogowski in
Current site
Google Scholar
PubMed
Close
https://orcid.org/0000-0003-0225-5341
,
Mark Otero Coastal Observing Research and Development Center, Marine Physical Laboratory, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, California

Search for other papers by Mark Otero in
Current site
Google Scholar
PubMed
Close
,
Joel Hazard Coastal Observing Research and Development Center, Marine Physical Laboratory, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, California

Search for other papers by Joel Hazard in
Current site
Google Scholar
PubMed
Close
,
Thomas Muschamp Coastal Observing Research and Development Center, Marine Physical Laboratory, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, California

Search for other papers by Thomas Muschamp in
Current site
Google Scholar
PubMed
Close
,
Scott Katz Coastal Observing Research and Development Center, Marine Physical Laboratory, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, California

Search for other papers by Scott Katz in
Current site
Google Scholar
PubMed
Close
, and
Eric Terrill Coastal Observing Research and Development Center, Marine Physical Laboratory, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, California

Search for other papers by Eric Terrill in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Accurate surface meteorological (MET) observations reported reliably and in near–real time remain a critical component of on-scene environmental observation systems. Presented is a system developed by Scripps Institution of Oceanography that allows for rapid, global deployment of ground-based weather observations to support both timely decision-making and collection of high-quality weather time series for science or military applications in austere environments. Named the Expeditionary Meteorological (XMET), these weather stations have been deployed in extreme conditions devoid of infrastructure ranging from tropical, polar, maritime, and desert environments where near continuous observations were reported. To date, over 1 million weather observations have been collected during 225 deployments around the world with a data report success rate of 99.5%. XMET had its genesis during Operation Iraqi Freedom (OIF), when the U.S. Marine Corps 3rd Marine Aircraft Wing identified an immediate capability gap in environmental monitoring of their operation area due to high spatiotemporal variability of dust storms in the region. To address the sensing gap, XMET was developed to be a portable, expendable, ruggedized, self-contained, bidirectional, weather observation station that can be quickly deployed anywhere in the world to autonomously sample and report aviation weather observations. This paper provides an overview of the XMETs design, reliability in different environments, and examples of unique meteorological events that highlight both the unit’s reliability and ability to provide quality time series. The overview shows expeditionary MET sensing systems, such as the XMET, are able to provide long-term continuous observational records in remote and austere locations essential for regional spatiotemporal MET characterization.

ORCID: 0000-0003-0225-5341.

Supplemental information related to this paper is available at the Journals Online website: https://doi.org/10.1175/JTECH-D-20-0016.s1.

© 2021 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Peter Rogowski, progowski@ucsd.edu

Abstract

Accurate surface meteorological (MET) observations reported reliably and in near–real time remain a critical component of on-scene environmental observation systems. Presented is a system developed by Scripps Institution of Oceanography that allows for rapid, global deployment of ground-based weather observations to support both timely decision-making and collection of high-quality weather time series for science or military applications in austere environments. Named the Expeditionary Meteorological (XMET), these weather stations have been deployed in extreme conditions devoid of infrastructure ranging from tropical, polar, maritime, and desert environments where near continuous observations were reported. To date, over 1 million weather observations have been collected during 225 deployments around the world with a data report success rate of 99.5%. XMET had its genesis during Operation Iraqi Freedom (OIF), when the U.S. Marine Corps 3rd Marine Aircraft Wing identified an immediate capability gap in environmental monitoring of their operation area due to high spatiotemporal variability of dust storms in the region. To address the sensing gap, XMET was developed to be a portable, expendable, ruggedized, self-contained, bidirectional, weather observation station that can be quickly deployed anywhere in the world to autonomously sample and report aviation weather observations. This paper provides an overview of the XMETs design, reliability in different environments, and examples of unique meteorological events that highlight both the unit’s reliability and ability to provide quality time series. The overview shows expeditionary MET sensing systems, such as the XMET, are able to provide long-term continuous observational records in remote and austere locations essential for regional spatiotemporal MET characterization.

ORCID: 0000-0003-0225-5341.

Supplemental information related to this paper is available at the Journals Online website: https://doi.org/10.1175/JTECH-D-20-0016.s1.

© 2021 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Peter Rogowski, progowski@ucsd.edu

1. Introduction

Weather often plays a prominent role in the success of military operations (see Fuller 1974, 1990). Operation Iraqi Freedom (OIF) was originally scheduled to commence on 12 March 2003 but, due to adverse weather forecasts and uncertainty in the impact to operations, it was delayed by a week to allow for further assessment. A second storm followed stalling thousands of U.S. troops and military vehicles stationed at Kuwait’s northern borders from advancing into Iraq. While waiting the storm out, damage from blowing sand and dust that penetrated engines and electronics caused further delays (Saeed and Al-Dashti 2011). Challenges continued in Afghanistan during Operation Enduring Freedom (OEF) where extreme “brown-out” events often reduced visibility to tens of meters off the ground (Miller 2003). Dust storm forecasting and tracking of the ubiquitous dust sources characteristic of southeast Asia during these campaigns were of paramount importance to military planners. The seasonal variation of dust activity in the Middle East is complex and regionally dependent. Over much of the Middle East, dust is active all year long, with increased activity starting in the spring and peaking in the summer (Shao 2008). Generally, spring dust storms are easier to forecast and track due to winds that are generated by the large-scale interactions of contrasting air masses over the Mediterranean. However, observing and predicting summer dust storms is inherently more difficult since they are typically generated by strong northwesterly winds that can last weeks and periodically intensify to gale force conditions (i.e., >17.5 m s−1). The interaction of these intense winds with dust source regions cause rapid, highly localized upwellings of hot air that generate dust plumes (Reynolds 2002; Vishkaee et al. 2011, 2012; Hamidi et al. 2013). In addition, collapsing summer thunderstorms and numerous dust source regions also generate a weather phenomenon characterized by immense walls of blowing sand and dust known in Arabic as a “haboob,” meaning “strong wind” (Fig. 1; Miller et al. 2008). Joint meteorological and oceanographic (METOC) personnel were tasked with forecasting and tracking these highly variable spatial and temporal events during OIF and OEF.

Fig. 1.
Fig. 1.

A massive dust storm (haboob) envelops the military camp at Al Asad, Iraq, on 27 Apr 2005. Department of Defense (DoD) photo by Corporal Alicia M. Garcia, U.S. Marine Corps (released).

Citation: Journal of Atmospheric and Oceanic Technology 38, 1; 10.1175/JTECH-D-20-0016.1

Background, genesis of a new meteorological system

The U.S. Navy began active development of an automated weather station in 1939. The first operational system was produced in 1941 and weighed one ton (Wood 1946). As rapid advances in microelectronics, computers, and communication technologies occurred toward the end of the twentieth century, deployments of automated weather observing stations became more practical resulting in widespread usage by the federal government (Fiebrich 2009). Today, the National Oceanic Atmospheric Administration’s (NOAA) Automated Surface Observing System (ASOS) serves as a primary climatological observing network in the United States. Hundreds of airports around the world utilize ASOS to support aviation operations and weather forecast activities. However, the systems are not practical for observing variable weather events in remote locations in a rapidly evolving expeditionary campaign due to infrastructure requirements and costs/time associated with installation and maintenance of a network of systems. As a result, weather observers during OIF and OEF lacked in situ data needed to characterize environmental conditions in areas of operation, and required human observers to be stationed at diverse sets of forward operating bases (FOBs). Adding to the challenge were limitations in available remote sensing methods for detecting and characterizing dust plumes at the appropriate space and time scales. Satellite radiometers, such as those employing multispectral dust enhancement techniques (e.g., Miller 2003), are limited to daytime cloud-free passes making it difficult to consistently detect and map the evolution of intermittent and short-term dust events.

The U.S. Marine Corps (USMC) METOC support is an embedded capability tasked with supporting Marine Air–Ground Task Force (MAGTF) operations through the use of expeditionary METOC systems. Tasking includes in-field sensing, collection, assimilation, processing, dissemination, and integration of environmental data into METOC products and services. Their capabilities facilitate the dynamic characterization and understanding of both the current and future state of the operational area providing situational awareness for commanders and warfighters during planning and battlespace operations (Joint Staff 2018).

During OIF, area weather observations for the Al Anbar province in Iraq were provided by the USMC METOC forces who relied on Marine weather personnel and observers deployed at air bases and FOBs. The logistics of these deployments (e.g., a minimum of two Marines deployed to each weather observation site) constrained the number of weather observation sites due to manning limitations. The lack of a robust observation network inhibited weather forecasters from providing complete situational awareness to battlespace commanders which ultimately impacted aviation and ground operations. Extreme drought conditions prevalent throughout the Middle East in 2008 (see Barlow et al. 2016) exacerbated dusty conditions during OIF resulting in a significant increase in the number of dust events that degraded the MAGTF’s ability to conduct operations throughout the Al Anbar province. To overcome the immediate capability gap in weather observations and environmental sensing awareness within their operational area, the USMC 3rd Marine Aircraft Wing (MAW) collaborated with Scripps Institution of Oceanography (SIO) and the U.S. Office of Naval Research (ONR) to develop a portable, rapidly deployable, self-contained, and autonomous environmental sensor capable transmitting standard aviation meteorological (MET) parameters. Requirements included a portable self-contained power system (e.g., solar power) and satellite communication allowing for long-term continuous deployment in remote locations with minimal maintenance. In addition, the cost of the system needed to be low so that multiple sensor systems could be deployed to create a network of weather observations at the growing number of FOBs occupied by the USMC. Initiated in 2008, an autonomous Expeditionary Meteorological (XMET) sensor system was rapidly prototyped and fielded to the USMC to provide critical feedback in refining and hardening the technology.

In this paper, we provide an overview of the XMETs design, operation, capabilities, and some example datasets highlighting both the survivability and ability to provide measurements from demanding environments for a variety of applications. In the following section, a technical overview of the current operational system including hardware, instrumentation, communication, data processing and dissemination is provided. In section 3 an overview of XMET performance, expanded applications, lessons learned from a range of operating environments, and example datasets of extreme atmospheric events including dust storms and typhoons are presented.

2. Expeditionary Meteorological station design

a. Hardware design

The XMET was designed to be modular using mistake-proof connectors and minimal components that can be assembled without the use of tools (Fig. 2b). Rust-proof hardware and anodized components were selected to endure the elements while electronics were protected by mil-spec waterproof connectors a watertight junction box. Once familiar with the system, one person can typically deploy the XMET in under 5 min while two people can do it in under 2 min. An optional ceilometer, for measuring cloud cover and ceiling heights can be added to the system.

Fig. 2.
Fig. 2.

(a) Image of modular XMET components packed in a ruggedized shipping case. (b) Pictorial illustrations showing assembly instructions of XMET components. (c) Computer-aided design (CAD) image of assembled XMET system with optional ceilometer.

Citation: Journal of Atmospheric and Oceanic Technology 38, 1; 10.1175/JTECH-D-20-0016.1

While the overall design of the XMET has largely remained unchanged throughout its development, nearly every component has been improved based on field experience and failure modes. The tripod at the base of the unit was originally a commercial collapsible product designed to hold sound equipment. While the tripod was sturdy, light, and relatively inexpensive, it became difficult to collapse and reuse once moving parts were exposed to dirt, sand, or pebbles. As a result, we designed a stronger, lightweight custom tripod that was easy to secure and remained usable over several redeployments.

A Hoffman electronics enclosure was originally used for the junction box which houses all the electronics and battery. Although the enclosure selected had a NEMA rating of 4X, it did not stand up to the rigors of remote deployments for long duration. We observed leaks at the lid seal which allowed moisture to accumulate on electronics, occasionally resulting in failure. Similar issues were observed on nonairtight welds. To address these issues, a custom junction box was fabricated from 6061 aluminum sheet metal that also included a custom lid with a SAE boss plug used to purge the box with nitrogen removing internal condensation prior to deployment. The original system was also updated with a more robust connector, the Souriau UTS Hi Seal MIL-DTL-26482 series that provided a seal even if the connector was not fully mated to the bulkhead surface on the junction box. The final design has been observed to remain airtight in the field and even waterproof in the case of a unit that was buried in coral rubble during a typhoon.

The XMET is packed into a single ruggedized SKB Case R Series 4024-18 shipping case to ensure all parts remain together when being fielded. This required designing a custom closed cell foam to layer components securely in place during transit. The case dimensions are 0.5 m × 0.7 m × 1.1 m and, once loaded with an XMET, has a shipping weight of 55 kg (Fig. 2a) which is suitable for direct commercial or military air transport and can be hand carried by two people.

b. Instrumentation, electronics, communication, and onboard processing

The XMET is an entirely self-contained unit that does not require power or communication infrastructure for deployment. Industry standard commercial off-the-shelf weather sensors were selected to ensure both robust instrument design and high data quality that meets standards set by the International Civil Aviation Organization (ICAO), in close collaboration with the World Meteorological Organization (WMO). Power is provided by a solar panel and rechargeable battery while an integrated satellite modem provides global bidirectional communications.

A low-power microcontroller (Persistor CF2) is programmed to execute the XMETs sampling duty cycle, which includes performing data acquisition and satellite telemetry. The CF2 clock is synchronized prior to shipping but clock drift resulted in a drifting duty cycle when units were deployed for extended periods. Software was developed to synchronize the system clock to GPS time with each duty cycle which eliminated the drift issue. In the event of GPS failure, the satellite telemetry message time proved to be a reliable independent clock for assigning a time stamp to reported data values.

Iridium’s polar-orbiting constellation provides global coverage and the low-latency short burst data (SBD) service supports telemetry of up to 340 bytes per message. Each MET report is encoded, along with system engineering information, into a single SBD that is configured to deliver over email (Fig. 3). The SBD service was relatively new when the XMET was being developed and connections between the modem and satellite constellation were unstable. By implementing a retry routine, with a maximum of three attempts, we were able to achieve over 99% message throughput. A summary of the XMET components including function, benefits and limitations is provided in Table 1.

Fig. 3.
Fig. 3.

XMET systems communicate through Iridium’s bidirectional SBD protocol and a local mail server. Data are processed immediately upon delivery and made available through a variety of services in order to accommodate several endpoints.

Citation: Journal of Atmospheric and Oceanic Technology 38, 1; 10.1175/JTECH-D-20-0016.1

Table 1.

XMET component description and overview.

Table 1.

With the exception of the rain sensor on the WXT530, which is always powered in order to continuously integrate rainfall and hail measurements, the XMET remains in a sleep state to conserve power. Ten minutes prior to its scheduled reporting time, in accordance with the U.S. Office of the Federal Coordinator for Meteorological Services (OFCM) Federal Meteorological Handbook 1 (OFCM 2017) the system wakes the GPS, compass and all weather sensors and collects streaming data in memory. After the 10-min collection period, data are averaged and prepared for transmission, the rain sensor is reset, raw data are stored to the CF card and the data are sent via SBD. When a ceilometer is powered by the system, the sampling window provides enough integration time for three levels of cloud-base heights or vertical visibility. The XMET is typically configured to transmit hourly reports of weather observations at 55 min past the hour in accordance with WMO directives. However, if the weather is rapidly changing or there is an operational need for additional reports, the XMET can be remotely commanded to increase sampling to twice an hour, at 25 and 55 min past the hour.

Earlier versions of the XMET used an Airmar PB200 for making meteorological observations. While it had several appealing features including no moving parts and an onboard compass and accelerometer for measuring true wind, it was easily fouled in nonmaritime environments and had a high power consumption. The WXT520 provided more accurate observations and used less than 5% of the total power consumed by the Airmar PB200 (~3.5 mA at 12 VDC). Similarly, initial XMET systems contained the Sentry Visibility Sensor that was replaced by a Vaisala PWD20 Visibility sensor due to lower power consumption and a more favorable form factor.

As a result of over a decade of spiral development, the XMET electronics, sensors, and power system has been observed to last for years in the field with little to no maintenance. While regular sensor cleaning and calibration would be ideal in order to maintain measurement quality, it is often impractical in remote and hostile deployment locations. Optical sensors used for visibility and cloud ceiling heights require the most maintenance in order to keep the light path clear but even contaminated observations are often reliable as operational decision aids in the absence of any other observations or communications. In cases where little to no onsite maintenance was performed during deployments, valuable data were collected and shown to be consistent with either nearby manned observations stations or model output.

c. Server-side processing and data dissemination

The XMET sensor has an operational infrastructure that has matured to be scalable and flexible enough to support all stages of the sensor’s life cycle. A wide array of configurable parameters is enabled through bidirectional communications to customize unit operation and processing throughout its deployment as environmental and sampling conditions change. Engineers and software developers have access to several configuration options along with documentation and test environments for efficient management and problem solving. End users and government agencies are able to obtain information through several data service and interface options depending on the application and operational environment. All XMET versions developed over the last decade remain backward compatible with the current software architecture.

A full suite of command line executables as well as an engineering-level web interface facilitates configuration and management of the XMET fleet. Utilities provide interfaces for viewing logs, unit configuration, data, and status. Reliable and accurate log monitoring is at the core of many automated monitoring routines. Log monitoring is performed on all units and any anomalies (identified by severity) as well as any unit going offline or coming online triggers email notification to operators with embedded links to web sites to expedite response. A more detailed description of message processing and dissemination is provided in the online supplemental material.

3. Field results

The initial XMET systems developed for OIF provided hourly observations to Marine weather forecasters in Iraq that were used to monitor weather conditions at FOBs within the Al Anbar province. These observations enabled battlespace commanders to infuse their planning and decision-making process with accurate and time-sensitive environmental intelligence. During testing and validation at the Al Asad Air Base in July 2008, XMET observations confirmed spatially variable dust plumes originating from moderate wind conditions significantly impacted atmospheric visibility. Over the initial 50-day deployment period, visibility below 2 km occurred 7% of the time while visibility was below 1 km for 3% of the observation period (Fig. 4). The dust events often only lasted several hours, with many occurring at night, making them difficult to detect with satellite-based remote sensing products.

Fig. 4.
Fig. 4.

Time series of visibility (orange) and wind speed (gray) observations from an XMET deployed at the Al Asad Air Base in July 2008.

Citation: Journal of Atmospheric and Oceanic Technology 38, 1; 10.1175/JTECH-D-20-0016.1

The operational successes of the original XMET systems led to a request from the Department of Defense (DoD) for 11 additional units to create a near-real-time MET network in Afghanistan operation areas during OEF. The utilization of an array of systems provided weather forecasters with in situ observations at discrete locations of interest as well as an ad hoc network of continuous field observations that could be used to map and characterize regional atmospheric dynamics and validate forecasts without deploying additional personnel forward.

a. XMET performance

Fifty-four XMET sensor systems have been built since the start of the project in 2008. These systems have collected over 1 million meteorological observations from 225 global deployments (Fig. S1 in the online supplemental material). The majority of the deployments (approximately 80%) directly supported operational aviation in remote locations while the remainder provided MET observations in support of either operational weather observations and forecasting or research. Deployment lengths vary from a few days to several years with a median deployment length of nearly 2 months with most systems being relocated multiple times; a testament to their expeditionary nature.

The system’s reliability is expressed in terms of reporting success, defined as the number of weather observations received versus the number of expected reports based on continuous hourly reporting over the deployment length. Using this metric, we observe a median reporting success over all deployments of 99.5%. Twenty-five deployments exceed yearlong deployments and have a reporting success ranging from 96.3% to 99.98%. Thirty-seven deployments had a reporting success of 100%. Over 90% of deployments exceeded 95% reporting success while the remaining 10% of deployments tended to be relatively short (less than 3 months) and may have been affected by poor electromagnetic conditions (nearby equipment or structures) resulting in intermittent data transmission. Although rare, the ultimate cause of a missed report is most commonly due to a failure in either satellite communication initiation or transmission.

The longest continuous unattended XMET deployment spanned over 6 years and 8 months (2443 days) from February 2011 until October 2017 (Fig. 5). The unit was deployed in western Afghanistan, collecting 58 145 weather observations with a reporting success of 99.2%. Two years into the deployment, the visibility sensor started issuing warnings of optical contamination that after another 8 months without maintenance was elevated to an alarm signaling reliable measurements were no longer possible. However, continuous observations were collected by the remaining sensors throughout the deployment capturing seasonal trends in air temperature ranging between 0° and 45°C and peaks in wind speed corresponding to strong northwesterly summer winds (Figs. 5a,b). Summer winds are accompanied by the monsoon trough that sets up across southern Iran and the southern Arabian Peninsula (Wilderson 1991; Hamidi et al. 2013). The consistent seasonal setup of this low pressure system is evident in barometric pressure observations with minimum pressures occurring each year from June through August (Fig. 5c). From an engineering perspective, the seasonal signal evident in battery voltage, with a reduction of 1 V during the hot summer months (Fig. 5d), is a result of operational efficiency loss in the solar panel when temperatures exceed 25°C. The unit was not recovered, so the failure mode is not known, but observations suggest that the charging system ultimately failed (most likely due to a damaged solar panel based on other units deployed in the area) causing the system to run out of power.

Fig. 5.
Fig. 5.

Time series of (a) wind speed, (b) temperature, (c) barometric pressure, and (d) system voltage observations from an XMET deployed in western Afghanistan from 2011 February through October 2017 (2443 days). A total of 58 145 weather observations were collected with reporting success of 99.2%.

Citation: Journal of Atmospheric and Oceanic Technology 38, 1; 10.1175/JTECH-D-20-0016.1

b. Environmental regimes

The refinements and operational successes of XMET deployments in the Middle East provided credibility to a mobile system capable of reliably operating in rugged remote locations with no infrastructure. As a result, XMETs were applied to a wider set of environmental conditions/locations and applications that benefited from near-real-time MET observations (Fig. 6).

Fig. 6.
Fig. 6.

Global map showing XMET deployment locations since 2008. Deployment locations discussed in this paper are labeled.

Citation: Journal of Atmospheric and Oceanic Technology 38, 1; 10.1175/JTECH-D-20-0016.1

1) Maritime

XMET deployments along the archipelago of Palau provided long-term evaluations of the XMET systems in remote maritime environments. The first system (XM44) was deployed on the western barrier reef island of Ongingiang in Koror State in March 2012 (Figs. 7b and 8a) while a second unit (XM20) was installed at the northern end of the archipelago on Ngeruangl Island in October 2012 (Fig. 8a), with both islands only being a couple of meters above sea level. Palau is located at 7.5°N latitude in the northwest Pacific near the planet’s most active generation region for typhoons. While typhoons typically intensify at latitudes greater than 10°N (Gouezo et al. 2015; Chu et al. 2017), two El Niño–Southern Oscillation (ENSO)-driven typhoons, Bopha (December 2012) and Haiyan (November 2013), struck Palau while XMETs were deployed providing observations that were distributed to the Joint Typhoon Warning Center (JTWC) in near–real time.

Fig. 7.
Fig. 7.

XMETs deployed (a) in desert conditions, (b) in a maritime environment, (c) on a glacier, and (d) on sea ice.

Citation: Journal of Atmospheric and Oceanic Technology 38, 1; 10.1175/JTECH-D-20-0016.1

Fig. 8.
Fig. 8.

(a) Map of the primary islands in the Palau archipelago and the deployment locations of two XMET systems. (b) Map of the Vietnamese coastline with locations of three environmental installations denoted by the black triangles. An XMET system was integrated into each environmental site (see inset).

Citation: Journal of Atmospheric and Oceanic Technology 38, 1; 10.1175/JTECH-D-20-0016.1

Stand-alone XMET systems were primarily used in Palau due to a lack of infrastructure on the remote islands where they supported aviation and maritime law enforcement agencies. However, XMETs are capable of leveraging existing infrastructure while falling back on internal power and communications during external outages (e.g., storms). As part of a capacity building collaborative effort between ONR and U.S. and Vietnamese academic institutions (see Rogowski et al. 2019), three XMET systems were integrated into existing infrastructure used to house other environmental sensing equipment along the Vietnamese coast in 2016 (Fig. 8b). The environmental sites had unreliable power sources, particularly during storm events, making the self-contained XMET systems ideally suited for continuous MET observations in these remote regions. The systems were able to stay operational through Typhoon Doksuri which made landfall near the Dong Hoi station as a category 3 storm in September 2017 providing valuable observational datasets to our Vietnamese partners. Due to their destructive nature, field observations of these extreme weather systems remain sparse (Liu et al. 2015), making observations collected by expeditionary MET systems important for observing and forecasting typhoon paths and intensity.

2) Globemaster II recovery/Ice Exercise

In addition to capacity building and scientific applications, the XMET system continues to be utilized for military expeditionary operations that require near-real-time MET observations. In June of 2015 and 2016, an XMET was provided to support an annual collaborative effort between the U.S. Air Force, Alaska Army National Guard, and Defense Prisoner of War/Missing in Action Agency to recover wreckage and remains from a 1952 crash of a military transport aircraft at Colony Glacier on Alaska’s Mount Gannett. A C-124 Globemaster aircraft crashed on 22 November 1952 while en route to Elmendorf Air Force Base, Alaska, from McChord Air Force Base, Washington, with 51 passengers on board. The adverse weather conditions entombed the wreckage in snow and ice preventing search teams from locating and recovering lost service members. In June 2012, the wreckage was rediscovered at the foot of Colony Glacier allowing for seasonal recovery efforts during a small weather window every June (Duffie 2015). For the 2015 and 2016 recovery efforts, an XMET was deployed on Colony Glacier (Fig. 7c) to report MET conditions which were critical in determining safe weather windows for flight operations to and from the glacier.

More recently, XMET systems were provided to the Navy’s Arctic Submarine Laboratory (ASL) to support the multinational biennial Ice Exercise (ICEX), conducted since 1947, in the Arctic Ocean. The program serves to advance the understanding of the Arctic environment and as a proving ground for submarine Arctic operability and warfighting (Arctic Submarine Laboratory 2018). XMET systems were deployed during the 2016 and 2018 ICEX exercises to provide near-real-time MET observations for flight operations in and out of a remote ice camp in the Beaufort Sea (Fig. 7d), approximately 300 km from a logistics headquarters in Prudhoe Bay, Alaska. In addition to favorable weather conditions, flight operations also required smooth ice thick enough to support an aircraft during landing and takeoff, daylight, and stable ice floes, limiting the biennial exercise to a few weeks in late winter through early spring. However, the ice environment remains dynamic as evident during the 2016 exercise when rapid propagation of several ice leads forced an immediate evacuation of camp SARGO (Nelson 2016). Automated XMET observations during and after the emergency evacuation enabled controllers to conduct safe flight operations in and out of the compromised ice camp and allowed for recovery of additional equipment following the evacuation (Fig. S2).

c. Sensing of episodic events

1) Synoptic mapping

The high mountains in eastern Iran and southwestern Afghanistan steer winds through deep valleys and lowlands that leads to suspension and advection of dust. This region is also under the influence of the persistent northwesterly Levar winds that strengthen from mid-May to mid-September, a period known as the “wind of 120 days” (Goudie and Middleton 2006; Alizadeh-Choobari et al. 2014). The winds are intensified by a channeling effect as northwesterly winds flow between eastern mountain ranges in Iran and Herat mountains (Fig. 9) reaching speeds up to 15–20 m s−1. The strong winds flow over one of the most active sources of dust in southwest Asia (Middleton 1986; Goudie and Middleton 2006; Esmaili and Tajrishy 2006), the Sistan Basin containing the typically dry Hamoun Lakes. Upstream dams in addition to severe droughts over the last decades have caused desiccation of the lakes leaving a fine layer of loose sediment that is easily lifted by the wind (Ranjbar and Iranmanesh 2008). These dust sources along with ones located in the Dasht-e Margo Desert, are the generation sites for dust storms that affect southwestern Afghanistan (Kaskaoutis et al. 2014; Alizadeh-Choobari et al. 2014). Long-term (1963–2009) monthly averaged visibility observations taken at the city of Zabol (Fig. 9a), show a clear annual cycle with peak winds in June correlated to low visibility. Conversely, lower average winter wind speeds result in better visibility conditions (see Rashki et al. 2012). While summer conditions may be optimal for frequent and intense dust storms, interannual variability modulates the peak storm months.

Fig. 9.
Fig. 9.

Topographic map of southwestern Afghanistan with XMET deployment locations and wind observations marked by black squares and overlaid with 1-km resolution MODIS AOD products for (a) 31 and (c) 27 Jul 2010 and (b),(d) their corresponding objectively mapped visibility and wind estimates. Objectively mapped estimates that have ratios of error to signal variance smaller than 0.8 are plotted. The primary northwesterly wind direction (thick black arrow) and dust source regions (red dotted line) are denoted in (a). Two southern XMET sites in (a) are labeled (XM1, XM2) for discussion in the manuscript. Visibility contours derived from XMET observations are included for reference between MODIS and OA plots (black lines).

Citation: Journal of Atmospheric and Oceanic Technology 38, 1; 10.1175/JTECH-D-20-0016.1

Dust plumes generated in the Sistan Basin or Dasht-e Margo Desert can be advected across southern Afghanistan, south of the Herat Mountains by regional cyclonic circulation that can arise from intense surface heating, along with the development of a local low pressure system of thermal origin (Kaskaoutis et al. 2014). The topographically driven regional meteorology creates large spatiotemporal variability in dust events across southern Afghanistan making accurate mapping and forecasting a challenge. During OEF, an ad hoc network of XMET sensors were deployed along the northern and eastern boundaries of the Dasht-e Margo Desert for improved situational awareness during ground and aviation operations. The two southern XMET sites experienced the largest number of low visibility events (<2 km), consistent with expected cyclonic circulation patterns that advects dust plumes from their source regions (e.g., Figs. 9a,c). The annual trends from these southern sites were similar to observations from Zabol where higher winds correspond to lower visibility during the summer months (Fig. S3a).

Large spatial variability in visibility during dust storms is evident between XMET observations at neighboring sites. A comparison of visibility observations between the two southernmost XMET sites (XM1 and XM2; Fig. 9a) for a 2-week period in 2009 shows significantly fewer low-visibility events for XM1 than for the XM2 site 64 km to the north (Fig. S3b). The inhomogeneous atmospheric conditions were captured by an ad hoc network of XMET systems which are leveraged to create visibility and wind maps that represent synoptic scales and capture their evolution.

Objective analysis (OA), a statistical estimator based on the Gauss–Markov theorem (Bretherton et al. 1976), was used to map wind velocity and visibility observations from in situ observations as a proof of concept for a near-real-time product from a network of XMETs. OA is widely used in oceanic and atmospheric sciences to map spatially nonuniform data to a regularly spaced set of gridded values (Daley 1991; Emery and Thomson 2001). The resulting objective map is the minimum mean-square error estimate of a continuous function of a variable, given discrete data. A more detailed discussion of the OA technique applied to oceanographic data can be found in Bretherton et al. (1976) and Davis (1985).

Afghanistan’s topography and local dynamics produce high spatiotemporal variability leading to relatively short decorrelation lengths scales in comparison to the XMET station spacing that typically ranged from 50 to 100 km (Fig. S4). Performing OA on the ad hoc XMET network provides a test case for the design of future networks. When feasible, the inexpensiveness and mobility of the systems allows for network design adjustments and optimization based the configuration’s ability to resolve features of interest. An estimated decorrelation length scale from the Afghanistan network of 50 km suggests additional XMETs would need to be deployed to improve visibility/wind estimates between sites. However, OA on the ad hoc network illustrates how synoptic-scale features can be mapped to provide large-scale situational awareness in regions that experience complex meteorology despite suboptimal site locations.

We illustrate the utility of these OA products by comparing them to aerosol optical depth (AOD) products from the Moderate Resolution Imaging Spectroradiometer (MODIS) on NASA’s Terra and Aqua satellites. We use the standard MODIS Collection 6 Multi-Angle Implementation Atmospheric Correction (MAIAC) product (MCD19A2) that provides a suite of atmospheric products at 1 km spatial resolution including AOD (see Lyapustin et al. 2011). Since mineral dust is a key component of aerosol, satellite observations of AOD can be utilized as a proxy to map dust storms (e.g., Prasad and Singh 2007; Ridley et al. 2013).

Comparisons of AOD products to OA wind and visibility maps for two dust events on 31 July 2010 and 27 July 2010 illustrate the topographically influenced meteorology for the region (Fig. 9). The MODIS images for both events show cyclonic circulation that advects dust from the source regions toward XMETs deployed to the east (Figs. 9a,c). Observed southeasterly wind directions from XM1 and XM2 are consistent with cyclonic flow on the eastern flank of the dust storms. Spatial gradients in visibility from XMET observations also compare well with the dust storm boundaries as defined by the MODIS images (Figs. 9b,d). The large u-shaped cyclonic storm on 27 June 2010 likely formed a haboob (e.g., Fig. 1) that engulfed stations XM1 and XM2, corresponding to visibility observations near zero, while good visibility conditions were observed to the north. OA maps derived from the XMET array captured some of the patchiness in the eastern side of the dust plume delineating potential go–no go regions for ground or aviation operations. The analysis suggests that additional XMETs deployed further upstream, closer to the source dust regions, could improve forecasts of dust storms at the southern XMET sites.

2) Typhoon Bopha/Haiyan (Palau, 2012/13)

Since 2012, XMET systems deployed in maritime locations have been impacted by several typhoons providing insight into MET conditions during these intense storm events. One important scientific application of these datasets include assessing and improving the skill of wave models that have shown to be dependent on the quality of the input wind fields (Teixeira et al. 1995; Holthuijsen et al. 1996; Ponce de León and OcampoTorres 1998).

Western Pacific typhoons generally pass Palau to the North having minimal impact on the archipelago, yet two super typhoons, Bopha on 2 December 2012 and Haiyan on 6 November 2013, made near direct strikes on the island nation. Bopha passed 40 km south of the southern island of Anguar as a strengthening category 4 typhoon (Fig. 10a), with maximum sustained winds of 70 m s−1 (Chu et al. 2017; Gouezo et al. 2015). Due to the southern trajectory of the typhoon in relation to Palau, the eastern side of Palau’s largest island (Babeldaob), the southernmost islands of the main Palau group (Angaur, Peleleiu), and the Rock Island area were most impacted by the strong winds along the northeast quadrant of the cyclonic storm. The multigrid spectral wave model Wavewatch III (Tolman 1999), with operational National Centers for Environmental Prediction (NCEP) winds as input forcing, predicted significant wave heights of 9–12 m that, in combination with the typhoons storm surge, inundated the eastern side of central Babeldaob Island. The flooding caused widespread damage to homes and other structures in low-lying villages and massive destruction to reefs on the eastern island slope (Coral Reef Research Foundation 2014). Yet, despite being deployed on a small island just above sea level (Fig. 7b), the location of XM44, in the lee of the storm on the western side of the Palau archipelago (Fig. 10a), protected it from the typhoon’s storm surge resulting in MET observations that detailed the variability of the storm as it passed Palau to the south. The cyclonic storm rotation evident from NCEP wind fields (Fig. 10a, inset) was also apparent in XMET observed winds. As the typhoon approached Palau, northeasterly winds transitioned to easterly during the period of maximum intensity, then shifted to southeasterly as the storm passed, patterns consistent with the trajectory and rotation of the typhoon (Figs. 11a,b). A similar pattern, but with weaker winds, was observed at the second XMET system on Ngeruangel Island (XM20) located approximately 190 km north of Bopha’s trajectory. The minimum barometric pressure measured at XM44 as the typhoon passed to the south of Palau was 750.3 mmHg (1 mmHg ≈ 1.33 hPa).

Fig. 10.
Fig. 10.

Map of the western Pacific overlaid with the trajectories of Typhoons (a) Bopha and (b) Haiyan. The dotted black box denotes a region illustrated within the inset showing the wind circulation pattern of each storm as it approached two XMET locations (XM20 and XM44). Wind quiver arrows are color coded by intensity. Black quiver arrows at XM20 and XM44 denote speed and direction of observed XMET winds. Typhoon trajectories and maximum sustained wind speeds are provided by the JTWC unless otherwise noted, and were downloaded from https://www.metoc.navy.mil/jtwc/jtwc.html?best-tracks/.

Citation: Journal of Atmospheric and Oceanic Technology 38, 1; 10.1175/JTECH-D-20-0016.1

Fig. 11.
Fig. 11.

Palau XMET time series of observed (a),(d) maximum wind speed, (b),(e) direction, and (c),(f) barometric pressure during the passage of Typhoons (a)–(c) Bopha and (d)–(f) Haiyan.

Citation: Journal of Atmospheric and Oceanic Technology 38, 1; 10.1175/JTECH-D-20-0016.1

Eleven months after Typhoon Bopha, the eye of Super Typhoon Haiyan passed within two miles of XM20, deployed on Ngeruangel Island (Fig. 10b), causing significant damage to the island of Kayangel and northern Babeldaob states. Compared with Typhoon Bopha, Haiyan was a stronger category 5 as it passed the Palau archipelago on 6 November 2013, with estimated sustained winds of 80 m s−1 and a central pressure of 679.5 mmHg (Chu et al. 2017; Gouezo et al. 2015). XM20 went offline at 1500 UTC with observed winds out of the north at 33 m s−1, and a last barometric pressure observation of 736.6 mmHg. The typhoon’s eye was 70 km east of Ngeruangel (Fig. 10b) with forecasted significant wave heights of 9–11 m in the region (from Wavewatch III). The high sea state coupled with storm surge caused by the approaching typhoon eye inundated Ngeruangel Island and XM20. A site survey by the Coral Reef Research Foundation post–Typhoon Haiyan confirmed significant overtopping of the island had occurred removing all recognizable landmarks on the island including a stone cairn (Fig. S5a). The XMET system was found upside down 50 m from its original location, partially submerged during high tide and buried under up to 2 m of coral rubble (Fig. S5b). Despite being submerged for a week, destroying the external instruments (e.g., visibility, MET, solar panel), the interior of the electronics box remained dry and undamaged allowing for rapid redeployment (P. Colin, Coral Reef Research Foundation, 2020, personal communications). The XMET system located southwest of Babeldaob (XM44), was approximately 145 km south of the Haiyan track and measured maximum winds of 26 m s−1, with wind directions consistent with the trajectory and cyclonic rotation of the storm as it passed Palau to the north (Figs. 11d–f).

A comparison of October 2012–November 2013 Palau XMET wind observations to NCEP modeled winds show good correlation (r2 = 0.81) but modeled winds underestimate observed speeds (Fig. S6). Spectral analysis of wind speed shows modeled values underestimate energy at frequencies above 0.5 cpd (Figs. S5a,b). Model speeds capture 40% of the total variance observed in situ at XM20. Approximately 30% of the total in situ variance is beyond the model’s temporal resolution (4 cpd). However, even when accounting for the discrepancy in temporal resolution and integrating over the same limits, only 57% of the total variance observed at XM20 is captured by the model. Similar spectral energy deviations in wind speed occur for comparisons between the model and XM44 observations (Fig. S7b) with an additional bias in modeled wind direction due to a failure to capture topographic steering on the leeward side of the Palau islands (Fig. 8a). The region has prevailing northeast trade winds from November to June and southwest or northwest winds from July to October (Kayanne et al. 2002) that are well captured by XM20 and collocated simulated NCEP winds (Figs. S7c,e). However, winds around Palau vary with the topography of the islands as exhibited by more northeasterly winds (parallel to Babeldaob Island) at XM44 that are not captured by the collocated NCEP modeled winds (Figs. S5d,f).

4. Conclusions

Collecting continuous MET observations from remote regions remains a challenge due to the lack of existing infrastructure for power and communications as well as the costs and logistics associated with maintenance. Here we describe the origins, components, and applications of a self-contained expeditionary (i.e., portable) meteorological sensor system (XMET) designed for long-term operation in remote regions with no maintenance. The system disseminates near-real-time standard aviation MET parameters, including visibility, that are essential for continuous situational awareness in demanding environments.

The temporal lag in observations from satellite remote sensing products and differences between XMET observations and modeling products in remote environments show the importance of in situ observations during extreme weather events (e.g., dust storms, typhoons). Remotely sensing dust storms or modeling the complex wind fields exhibited by these storms is challenging due to the strong spatial and temporal gradients. The expeditionary nature of systems such as the XMET that can be rapidly deployed ahead of extreme weather can characterize these often poorly resolved events through persistent in situ MET observations that can be applied to model validation and training.

Since the origin of the XMET program in 2009, 54 systems have been fabricated and deployed over 225 times around the world with deployment lengths ranging from days to years; a testament to their expeditionary nature. Regardless of the environment (e.g., desert, polar, glacier, tropics), these systems have a proven reliability record with an average reporting success rate of 99.5% with over one million observations. Operational successes of the XMET systems proves that expeditionary MET sensor systems can be deployed in poorly constrained MET environments to create a network that can be easily reconfigured based on sampling objectives. These expeditionary systems provide reliable cost-effective solutions to operational and scientific applications where MET sensing of remote and extreme environments is required.

Acknowledgments

This paper is dedicated to SMSgt Ronald Kellerman USAF (ret.)—an ever humble hero providing tactical weather support over a career spanning 35 years of active duty and civilian service prior to his tragic passing. The authors wish to acknowledge the vision, leadership, and sponsorship provided by the U.S. Office of Naval Research, Code 32 (Ocean Battlespace and Expeditionary Access), which has supported the development and maturation of the XMET system, and, in particular, we thank program managers Thomas Drake, Daniel Eleuterio, Scott Harper, and Theresa Paluszkiewicz. We recognize the USMC 3rd Marine Aircraft Wing for their program support. Additionally, we recognize the Palau national government, Bureau of Marine Resources, Koror and Kayangel state governments, and Conservation officers for program support. We also thank the Coral Reef Research Foundation, in particular Patrick and Lori Colin, Mathew Mesubed, and Emilio Basilius, whose support was essential to the success of the program. Data used in this paper are available on figshare.

REFERENCES

  • Alizadeh-Choobari, O., P. Zawar-Reza, and A. Sturman, 2014: The “wind of 120 days” and dust storm activity over the Sistan Basin. Atmos. Res., 143, 328341, https://doi.org/10.1016/j.atmosres.2014.02.001.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Arctic Submarine Laboratory, 2018: ICEX 2018 briefing book. Arctic Submarine Laboratory Doc., 29 pp.

  • Barlow, M., B. Zaitchik, S. Paz, E. Black, J. Evans, and A. Hoell, 2016: A review of drought in the Middle East and southwest Asia. J. Climate, 29, 85478574, https://doi.org/10.1175/JCLI-D-13-00692.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bretherton, F. P., R. E. Davis, and C. B. Fandry, 1976: A technique for objective analysis and design of oceanographic experiments applied to MODE-73. Deep-Sea Res. Oceanogr. Abstr., 23, 559582, https://doi.org/10.1016/0011-7471(76)90001-2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chu, J. H., A. Levine, S. Daida, D. Schiber, E. Fukuda, and C. R. Sampson, 2017: The Joint Typhoon Warning Center tropical cyclone best-tracks, 1945–2017. Naval Meteorology and Oceanography Command, accessed 1 March 2020, https://www.metoc.navy.mil/jtwc/jtwc.html?best-tracks.

  • Coral Reef Research Foundation, 2014: A summary of Palau’s typhoon history 1945-2013. Coral Reef Research Foundation Tech. Rep., 17 pp.

  • Daley, R., 1991: Atmospheric Data Analysis. Vol. 2. Cambridge University Press, 471 pp.

  • Davis, R. E., 1985: Objective mapping by least squares fitting. J. Geophys. Res., 90, 47734777, https://doi.org/10.1029/JC090iC03p04773.

  • Duffie, W., 2015: ONR-sponsored technology aids recovery of Alaska plane wreck. ONR, https://www.onr.navy.mil/en/Media-Center/Press-Releases/2015/XMET-Used-In-Alaska-Plane-Crash-Recovery.

  • Emery, W. J., and R. J. Thomson, 2001: Data Analysis Methods in Physical Oceanography. Gulf Professional Publishing, 638 pp.

  • Esmaili, O., and M. Tajrishy, 2006: Results of the 50 year ground-based measurements in comparison with satellite remote sensing of two prominent dust emission sources located in Iran. Proc. SPIE, 6362, 636209, https://doi.org/10.1117/12.692989.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fiebrich, C. A., 2009: History of surface weather observations in the United States. Earth-Sci. Rev., 93, 7784, https://doi.org/10.1016/j.earscirev.2009.01.001.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fuller, J. F., 1974: Weather and war. U.S. Air Force Air Weather Service Rep., 25 pp.

  • Fuller, J. F., 1990: Thor’s Legions. Amer. Meteor. Soc., 443 pp.

  • Goudie, A. S., and N. J. Middleton, 2006: Desert Dust in the Global System. Springer, 287 pp.

  • Gouezo, M., Y. Golbuu, R. van Woesik, L. Rehm, S. Koshiba, and C. Doropoulos, 2015: Impact of two sequential super typhoons on coral reef communities in Palau. Mar. Ecol. Prog. Ser., 540, 7385, https://doi.org/10.3354/meps11518.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hamidi, M., M. R. Kavianpour, and Y. Shao, 2013: Synoptic analysis of dust storms in the Middle East. Asia-Pac. J. Atmos. Sci., 49, 279286, https://doi.org/10.1007/s13143-013-0027-9.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Holthuijsen, L. H., N. Booji, and L. Bertotti, 1996: The propagation of wind errors through ocean wave hindcasts. J. Offshore Mech. Arct. Eng., 118, 184189, https://doi.org/10.1115/1.2828832.

    • Search Google Scholar
    • Export Citation
  • Joint Staff, 2018: Meteorological and oceanographic operations. U.S. Joint Chiefs of Staff Publ. 3-59, 90 pp.

  • Kaskaoutis, D., A. Rashki, E. Houssos, A. Mofidi, D. Goto, A. Bartzokas, P. Francois, and M. Legrand, 2014: Meteorological aspects associated with dust storms in the Sistan region, southeastern Iran. Climate Dyn., 45, 407424, https://doi.org/10.1007/s00382-014-2208-3.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kayanne, H., H. Yamano, and R. H. Randall, 2002: Holocene sea-level changes and barrier reef formation on an ocean island, Palau Islands, western Pacific. Sediment. Geol., 150, 4760, https://doi.org/10.1016/S0037-0738(01)00267-6.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Liu, C. C., T. Y. Shyu, T. H. Lin, and C. Y. Liu, 2015: Satellite-derived normalized difference convection index for typhoon observations. J. Appl. Remote Sens., 9, 096074, https://doi.org/10.1117/1.JRS.9.096074.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lyapustin, A., Y. Wang, I. Laszlo, R. Kahn, S. Korkin, L. Remer, and J. S. Reid, 2011: Multiangle Implementation of Atmospheric Correction (MAIAC): 2. Aerosol algorithm. J. Geophys. Res., 116, D03211, https://doi.org/10.1029/2010JD014986.

    • Search Google Scholar
    • Export Citation
  • Middleton, N. J., 1986: Dust storms in the Middle East. J. Arid Environ., 10, 8396, https://doi.org/10.1016/S0140-1963(18)31249-7.

  • Miller, S. D., 2003: A consolidated technique for enhancing desert dust storms with MODIS. Geophys. Res. Lett., 30, 2071, https://doi.org/10.1029/2003GL018279.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Miller, S. D., A. P. Kuciauskas, M. Liu, Q. Ji, J. S. Reid, D. W. Breed, A. L. Walker, and A. A. Mandoos, 2008: Haboob dust storms of the southern Arabian Peninsula. J. Geophys. Res., 113, D01202, https://doi.org/10.1029/2007JD008550.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Nelson, M. S., 2016: Short-range acoustic propagation under Arctic ice cover ICEX-16. M.S. thesis, Dept. of Oceanography, Naval Postgraduate School, 55 pp.

  • OFCM, 2017: Surface weather observations and reports. OFCM Doc., 98 pp.

  • Ponce de León, S., and F. J. Ocampo-Torres, 1998: Sensitivity of a wave model to wind variability. J. Geophys. Res., 103, 31793201, https://doi.org/10.1029/97JC02328.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Prasad, A. K., and R. P. Singh, 2007: Changes in aerosol parameters during major dust storm events (2001–2005) over Indo-Gangetic plains using AERONET and MODIS data. J. Geophys. Res., 112, D09208, https://doi.org/10.1029/2006JD007778.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ranjbar, M., and F. Iranmanesh, 2008: Effects of “drought” on “wind eroding and erosion” in Sistan region with use of satellite multiple images. Islamic Azad Univ., Shahre Rey Branch, 14–30.

  • Rashki, A., and D. Kaskaoutis, C. J. de W. Rautenbach, and P., Eriksson, 2012: Changes of permanent lake surfaces, and their consequences for dust aerosols and air quality: The Hamoun Lakes of the Sistan area, Iran. Atmospheric Aerosols—Regional Characteristics—Chemistry and Physics, H. Addul-Razzak, Ed., InTech, 162–202, https://doi.org/10.5772/48776.

    • Crossref
    • Export Citation
  • Reynolds, R. M., 2002: Meteorology and climate. The Gulf Ecosystem: Health and Sustainability. N. Y. Khan, M. Munawar, and A. R. G. Price, Eds., Backhuys Publishers, 41–51.

    • Crossref
    • Export Citation
  • Ridley, D., C. Heald, J. Pierce, and M. J. Evans, 2013: Toward resolution-independent dust emissions in global models: Impacts on the seasonal and spatial distribution of dust. Geophys. Res. Lett., 40, 28732877, https://doi.org/10.1002/grl.50409.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rogowski, P., J. Zavala-Garay, K. Shearman, E. Terrill, J. Wilkin, and T. H. Lam, 2019: Air-sea-land forcing in the Gulf of Tonkin: Assessing seasonal variability using modern tools. Oceanography, 32 (2), 150161, https://doi.org/10.5670/oceanog.2019.223.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Saeed, T. M., and H. Al-Dashti, 2011: Optical and physical characterization of “Iraqi Freedom” dust storm, a case study. Theor. Appl. Climatol., 104, 123137, https://doi.org/10.1007/s00704-010-0334-3.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shao, Y., 2008: Physics and Modelling of Wind Erosion. Springer, 452 pp.

  • Teixeira, J. C., M. P. Abreu, and C. Guedes Soares, 1995: Uncertainty of ocean wave hindcasts due to wind modeling. J. Offshore Mech. Arctic Eng., 117, 294297, https://doi.org/10.1115/1.2827237.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tolman, H. L., 1999: User manual and system documentation of Wavewatch-III version 1.18. NOAA/NCEP Tech. Note 166, 110 pp.

  • Vishkaee, F. A., C. Flamant, J. Cuesta, P. Flamant, and H. R. Khalesifard, 2011: Multiplatform observations of dust vertical distribution during transport over northwest Iran in the summertime. J. Geophys. Res., 116, D05206, https://doi.org/10.1029/2010JD014573.

    • Search Google Scholar
    • Export Citation
  • Vishkaee, F. A., C. Flamant, J. Cuesta, L. Oolman, P. Flamant, and H. R. Khalesifard, 2012: Dust transport over Iraq and northwest Iran associated with winter shamal: A case study. J. Geophys. Res., 117, D03201, https://doi.org/10.1029/2011JD016339.

    • Search Google Scholar
    • Export Citation
  • Wilderson, W. D, 1991: Dust and sand forecasting in Iraq and adjoining countries. USAF Environmental Technical Applications Center Rep., 72 pp.

  • Wood, L. E., 1946: Automatic weather stations. J. Meteor., 3, 115121, https://doi.org/10.1175/1520-0469(1946)003<0115:AWS>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation

Supplementary Materials

Save