• Abramov, A. V., V. V. Rozhnov, and P. N. Morozov, 2006: Notes on mammals of the Ngoc Linh Nature Reserve (Vietnam, Kon Tum Province). Russ. J. Theriology, 2, 8592.

    • Search Google Scholar
    • Export Citation
  • Bengtsson, L., S. Hagemann, and K. I. Hodges, 2004: Can climate trends be calculated from reanalysis data? J. Geophys. Res., 109, D11111, doi:10.1029/2004JD004536.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bodenhamer, D. J., J. Corrigan, and T. M. Harris, 2010: The Spatial Humanities: GIS and the Future of Humanities Scholarship. Indiana University Press, 222 pp.

  • Buckland, M., A. Chen, F. C. Gey, R. R. Larson, R. Mostern, and V. Petras, 2007: Geographic search: Catalogs, gazetteers, and maps. Coll. Res. Libr., 68, 376387, doi:10.5860/crl.68.5.376.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Carbon Dioxide Information Analysis Center, Environmental Sciences Division, Oak Ridge National Laboratory, U. S. Department of Energy, Arizona State Climate Office/Arizona State University, and National Climatic Data Center/NESDIS/NOAA/U.S. Department of Commerce, 1995: Global Historical Climatology Net (GHCN) version 2 temperature, precipitation, pressure. Research Data Archive at the National Center for Atmospheric Research, Computational and Information Systems Laboratory, accessed 15 June 2017, http://rda.ucar.edu/datasets/ds564.0/.

  • Cavazos, T., 2000: Using self-organizing maps to investigate extreme climate events: An application to wintertime precipitation in the Balkans. J. Climate, 13, 17181732, doi:10.1175/1520-0442(2000)013<1718:USOMTI>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chang, C. P., Z. Wang, J. McBride, and C. H. Liu, 2005: Annual cycle of Southeast Asia—Maritime continent rainfall and the asymmetric monsoon transition. J. Climate, 18, 287301, doi:10.1175/JCLI-3257.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chen, T. C., J. D. Tsay, M. C. Yen, and J. Matsumoto, 2012: Interannual variation of the late fall rainfall in central Vietnam. J. Climate, 25, 392413, doi:10.1175/JCLI-D-11-00068.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chu, J. E., S. N. Hameed, and K. J. Ha, 2012: Nonlinear, intraseasonal phases of the East Asian summer monsoon: Extraction and analysis using self-organizing maps. J. Climate, 25, 69756988, doi:10.1175/JCLI-D-11-00512.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Compo, G. P., and Coauthors, 2011: The Twentieth Century Reanalysis Project. Quart. J. Roy. Meteor. Soc., 137, 128, doi:10.1002/qj.776.

  • Croizat, V. J., 1967: A translation from the French: Lessons of the war in Indochina. Vol. 2. Rand Corporation Memo. RM-5271-PR., 415 pp, http://www.rand.org/pubs/research_memoranda/RM5271.html.

  • Cromley, G., 2016: Designing a military event gazetteer: The case of parachute operations during the French Indochina War. Prof. Geogr., 68, 249260, doi:10.1080/00330124.2015.1062702.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • DAAF, 1967: U.S. Army/U.S. Air Force doctrine for airborne operations. Field Manual 57-1, Air Force Manual 2-51, 71 pp.

  • Davies, D. L., and D. W. Bouldin, 1979: A cluster separation measure. IEEE Trans. Pattern Anal. Mach. Intell., PAMI-1, 224227, doi:10.1109/TPAMI.1979.4766909.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davis, R. E., and L. S. Kalkstein, 1990: Development of an automated spatial synoptic climatological classification. Int. J. Climatol., 10, 769794, doi:10.1002/joc.3370100802.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dee, D., and Coauthors, Eds., 2016: The climate data guide: Atmospheric reanalysis: Overview & comparison tables. Accessed 15 June 2017, https://climatedataguide.ucar.edu/climate-data/atmospheric-reanalysis-overview-comparison-tables.

  • ECMWF, 2014: ERA-20C Project (ECMWF Atmospheric Reanalysis of the 20th Century). Research Data Archive at the National Center for Atmospheric Research, Computational and Information Systems Laboratory, accessed 15 June 2017, doi:10.5065/D6VQ30QG.

    • Crossref
    • Export Citation
  • Fall, B. B., 1966: Hell in a Very Small Place: The Siege of Dien Bien Phu. Da Capo Press, 568 pp.

  • Fall, B. B., 1994: Street Without Joy: The French Debacle in Indochina. Stackpole Books, 416 pp.

  • Ford, T. W., S. M. Quiring, O. W. Frauenfeld, and A. S. Rapp, 2015: Synoptic conditions related to soil moisture-atmosphere interactions and unorganized convection in Oklahoma. J. Geophys. Res. Atmos., 120, 11 51911 535, doi:10.1002/2015JD023975.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Goodchild, M. F., and L. L. Hill, 2008: Introduction to digital gazetteer research. Int. J. Geogr. Inf. Sci., 22, 10391044, doi:10.1080/13658810701850497.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Guttman, L., 1954: Some necessary conditions for common-factor analysis. Pyschometrika, 19, 149161, doi:10.1007/BF02289162.

  • Hewitson, B. C., and R. G. Crane, 2002: Self-organizing maps: Applications to synoptic climatology. Climate Res., 22, 1326, doi:10.3354/cr022013.

  • Hill, L. L., 2000: Core elements of digital gazetteers: Placenames, categories, and footprints. Research and Advanced Technology for Digital Libraries: Proc. Fourth European Conf., Lisbon, Portugal, ECDL, 280–290, doi:10.1007/3-540-45268-0_26.

    • Crossref
    • Export Citation
  • Huth, R., C. Beck, A. Philipp, M. Demuzere, Z. Ustrnul, M. Cahynová, J. Kyselý, and O. E. Tveito, 2008: Classifications of atmospheric circulation patterns: Recent advances and applications. Ann. N. Y. Acad. Sci., 1146, 105152, doi:10.1196/annals.1446.019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kaiser, H. F., 1960: The application of electronic computers to factor analysis. Educ. Psychol. Meas., 20, 141151, doi:10.1177/001316446002000116.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Keller, T., 2009: The mountains roar: The Alps during the Great War. Environ. Hist., 14, 253274, doi:10.1093/envhis/14.2.253.

  • Kohonen, T., 1995: Self-Organizing Maps. Springer-Verlag, 362 pp.

    • Crossref
    • Export Citation
  • Li, C., and M. Yanai, 1996: The onset and interannual variability of the Asian summer monsoon in relation to land–sea thermal contrast. J. Climate, 9, 358375, doi:10.1175/1520-0442(1996)009<0358:TOAIVO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Logevall, F., 2012: Embers of War: The Fall of an Empire and the Making of America’s Vietnam. Random House, 864 pp.

  • McCune, S., 1947: The diversity of Indochina’s physical geography. Far East. Quart., 6, 334344, doi:10.2307/2049430.

  • Meehl, G. A., 1987: The annual cycle and interannual variability in the tropical Pacific and Indian Ocean regions. Mon. Wea. Rev., 115, 2750, doi:10.1175/1520-0493(1987)115<0027:TACAIV>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mostern, R., 2011: “Dividing the Realm in Order to Govern”: The Spatial Organization of the Song State. Harvard University Press, 396 pp.

    • Crossref
    • Export Citation
  • Mostern, R., and I. Johnson, 2008: From named place to naming event: Creating gazetteers for history. Int. J. Geogr. Inf. Sci., 22, 10911108, doi:10.1080/13658810701851438.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • NOAA, 2013: French Indochina climatological data: Résumé mensuel des observations. Accessed 11 January 2017, https://www.lib.noaa.gov/collections/foreign_climate_data_pages/foreign_climate_data_indochina.html.

  • Pham, X. T., B. Fontaine, and N. Philippon, 2010: Onset of the summer monsoon over the southern Vietnam and its predictability. Theor. Appl. Climatol., 99, 105113, doi:10.1007/s00704-009-0115-z.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pissardy, J. P., 1982: Paras d’Indochine: 1944-1954. Société de Production Littéraire, 225 pp.

  • Poli, P., and Coauthors, 2013: The data assimilation system and initial performance evaluation of the ECMWF pilot reanalysis of the 20th-century assimilating surface observations only (ERA-20C). ERA Rep. Series 14, 62 pp., https://www.ecmwf.int/sites/default/files/elibrary/2013/11699-data-assimilation-system-and-initial-performance-evaluation-ecmwf-pilot-reanalysis-20th.pdf.

  • Poli, P., and Coauthors, 2016: ERA-20C: An atmospheric reanalysis of the twentieth century. J. Climate, 29, 40834097, doi:10.1175/JCLI-D-15-0556.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Porch, D., 2010: The French Foreign Legion: A Complete History of the Legendary Fighting Force. HarperCollins, 728 pp.

  • Sanderson, R. W., 1954: Notes on the climate of Indochina. Weatherwise, 7, 5669, doi:10.1080/00431672.1954.9930321.

  • Sheridan, S. C., and C. C. Lee, 2011: The self-organizing map in synoptic climatological research. Prog. Phys. Geogr., 35, 109119, doi:10.1177/0309133310397582.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shrader, C. R., 2015: A War of Logistics: Parachutes and Porters in Indochina, 1945–1954. University of Kentucky Press, 488 pp.

    • Crossref
    • Export Citation
  • Southall, H., R. Mostern, and M. L. Berman, 2011: On historical gazetteers. Int. J. Hum. Arts Comput., 5, 127145, doi:10.3366/ijhac.2011.0028.

    • Search Google Scholar
    • Export Citation
  • Takahashi, H. G., and T. Yasunari, 2006: A climatological monsoon break in rainfall over Indochina—A singularity in the seasonal march of the Asian summer monsoon. J. Climate, 19, 15451556, doi:10.1175/JCLI3724.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, B., and LinHo, 2002: Rainy season of the Asian–Pacific summer monsoon. J. Climate, 15, 386398, doi:10.1175/1520-0442(2002)015<0386:RSOTAP>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Winters, H. A., G. E. Galloway Jr., W. J. Reynolds, and D. W. Rhyne, 1998: Battling the Elements: Weather and Terrain in the Conduct of War. The Johns Hopkins University Press, 336 pp.

  • Wu, C. H., and H. H. Hsu, 2016: Role of the Indochina Peninsula narrow mountains in modulating the East Asian–western North Pacific summer monsoon. J. Climate, 29, 44454459, doi:10.1175/JCLI-D-15-0594.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yarnal, B., 1993: Synoptic Climatology in Environmental Analysis: A Primer. Belhaven Press, 256 pp.

  • Zhang, H., T. Li, B. Wang, and G. Wu, 2002: Onset of the summer monsoon over the Indochina Peninsula: Climatology and interannual variations. J. Climate, 15, 32063221, doi:10.1175/1520-0442(2002)015<3206:OOTSMO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery

    Map of the study area, including the selected reanalysis domain (dashed line), selected GHCN stations coded by data reporting percentage for the modern era (1981–2010), and the locations of all OAPs included in the Cromley (2016) gazetteer.

  • View in gallery

    Methodological framework for enriching a digital event gazetteer with weather data based on historical observations or reanalysis data. Note that the framework can be implemented for enriching digital gazetteers of other events using different weather variables.

  • View in gallery

    Daily mean sea level pressure patterns resulting from the SOM classification procedure. Darker blues indicate lower pressure, while darker reds indicate higher pressure. OAP locations (black dots) are plotted by the SOM pattern corresponding to each drop date.

  • View in gallery

    Graphs of average daily precipitation for each region (solid line) and theaterwide (dashed line) aggregated by month by region. The total number of OAPs during the conflict per month in that region is shown in bars. No OAPs occurred in Annam in November or December.

  • View in gallery

    Daily regional precipitation averages in Tonkin for all days associated with each day’s assigned SOM pattern for September 1951–February 1952. The inset graph depicts the same data for all days in 1951–52, with September–February highlighted. Stars indicate days on which one or more OAPs occurred during the Hoa Binh campaign in Tonkin.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 62 62 22
PDF Downloads 53 53 24

Assimilating Weather Data into a Digital Event Gazetteer of Airborne Parachute Operations during the French Indochina War

View More View Less
  • 1 Department of Geography, Kent State University, Kent, Ohio
© Get Permissions
Full access

Abstract

Modern datasets cataloging historical events, known as digital event gazetteers, feature spatiotemporal data regarding events that enable analysis through parameters including location and other descriptive information of those events. Weather and climate data represent two dimensions of spatiotemporal information, which can enhance understanding of historical events. A recently published digital event gazetteer of airborne parachute operations [opérations aéroportées (OAPs)] during and prior to the French Indochina War, spanning from 1945 to 1954, represents an opportunity to associate discrete historical events with weather information. This study outlines a methodology for assimilating weather data into the construct of a digital event gazetteer and then demonstrates example analyses of how the weather and climate conditions in Indochina may relate to OAPs during the war.

A synoptic classification, utilizing the self-organizing maps procedure, is performed using daily mean sea level pressure data from 1945 to 2010, from a twentieth-century reanalysis dataset, to characterize weather patterns over the Indochina Peninsula. Since observations are sparse during the years of the conflict, the resulting weather patterns are associated with modern precipitation observations in the area, as a representation of wet and dry patterns during the war. The appropriate daily weather pattern is then assigned to each OAP in order to investigate its relationship with the weather and climate patterns of Indochina, including the influence of monsoon seasons, and how the resulting precipitation patterns affected combat operations across the theater. Additionally, specific OAPs of various missions are analyzed to investigate how weather patterns may have affected operation planning during the French Indochina War.

© 2018 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: Johnathan P. Kirk, jkirk9@kent.edu

Abstract

Modern datasets cataloging historical events, known as digital event gazetteers, feature spatiotemporal data regarding events that enable analysis through parameters including location and other descriptive information of those events. Weather and climate data represent two dimensions of spatiotemporal information, which can enhance understanding of historical events. A recently published digital event gazetteer of airborne parachute operations [opérations aéroportées (OAPs)] during and prior to the French Indochina War, spanning from 1945 to 1954, represents an opportunity to associate discrete historical events with weather information. This study outlines a methodology for assimilating weather data into the construct of a digital event gazetteer and then demonstrates example analyses of how the weather and climate conditions in Indochina may relate to OAPs during the war.

A synoptic classification, utilizing the self-organizing maps procedure, is performed using daily mean sea level pressure data from 1945 to 2010, from a twentieth-century reanalysis dataset, to characterize weather patterns over the Indochina Peninsula. Since observations are sparse during the years of the conflict, the resulting weather patterns are associated with modern precipitation observations in the area, as a representation of wet and dry patterns during the war. The appropriate daily weather pattern is then assigned to each OAP in order to investigate its relationship with the weather and climate patterns of Indochina, including the influence of monsoon seasons, and how the resulting precipitation patterns affected combat operations across the theater. Additionally, specific OAPs of various missions are analyzed to investigate how weather patterns may have affected operation planning during the French Indochina War.

© 2018 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: Johnathan P. Kirk, jkirk9@kent.edu

1. Introduction

a. Digital gazetteer research

An emerging component of interdisciplinary research into historical events is the inclusion of weather and climate information. By assimilating weather and climate data into historical analysis, new research questions can be developed for investigating the relationships between historical or cultural phenomena and the physical environment within which they occurred. Spatial data have given researchers insight into historical narratives and provide greater context for the understanding of human history. A construct that has been recently developed for use in contemporary historical research is that of a digital event gazetteer. Digital gazetteers are defined as “geospatial dictionaries of geographic names” (Hill 2000, p. 280), with data entries composed into digital catalogs featuring at least three specific components: the name of a feature, the location of a feature, and some descriptive element that describes that feature (Goodchild and Hill 2008). Digital gazetteer research, in general, is a field that links places, which may be identified by numerous names, with discrete spatial data for the purposes of creating unique datasets that catalog places, as culturally defined entities, with spaces, the physical characteristics of each location (Buckland et al. 2007; Southall et al. 2011). Digital event gazetteers (Mostern and Johnson 2008) represent a specific variety of digital gazetteers, which catalog temporal events, providing a geographic location and other descriptive attributes for discrete events in time to compile an enriched database for the analysis of historical spatial data.

Using this form of geographic analysis, scholars have proposed methodologies for incorporating various aspects of geographic information science to enrich their own research (Bodenhamer et al. 2010). As spatial datasets, digital event gazetteers provide adaptable resources for data enrichment that have been utilized for a variety of applications, particularly among historical scholarship, such as understanding the changing spatial structures of governance in Song Dynasty China (Mostern 2011). Digital event gazetteers represent ideal resources for cataloging weather and climate data relevant to historical events to provide richer spatial databases of historical events for the benefit of subsequent analysis.

b. French Indochina War

One recent example of a digital event gazetteer cataloging specific events is a database documenting airborne operations from 1945 to 1954, during the French Indochina War and the events leading up to the conflict (Cromley 2016). The French Indochina War (1946–54) was a transitional conflict emerging out of the Pacific theater of World War II. The French government, which was trying to reoccupy their former colony, fought against a communist-led, nationalist movement under Ho Chi Minh called the Viet Minh. While there are numerous interpretations of the narrative of the conflict (e.g., Fall 1994; Logevall 2012), the conflict itself can be generally broken up into several thematic/temporal periods. The first (1945–46) is characterized by the French military expanding control over Indochina, first in Laos and then spreading out from Saigon in the south to retake control of Hanoi in the north. French forces in 1947 attempted to destroy the Viet Minh in their base area along the Chinese border. While this offensive failed, it did secure the majority of Tonkin in the north, from 1948 to 1950.

The end of 1949 marked a transition period when the communist victory in the Chinese Civil War provided the Viet Minh with a secure base area from which to support their forces. After the defeat of the French forces along the Chinese border area in 1950, the conflict transformed from an insurgent conflict into one that the French attempted to fight as a conventional conflict. The focus of the major military operations shifted almost exclusively to Tonkin and northern Laos. From 1951 until the end of the conflict, the French continually tried to force the Viet Minh forces into a massive battle similar to those in World War II. When it happened at Dien Bien Phu in 1954, despite the French’s technological advantages, the Viet Minh forces were able to defeat the French in a pitched battle and force France to negotiate a settlement to the conflict.

As in other armed conflicts (e.g., Winters et al. 1998; Keller 2009), weather and climate played an important role in the operational tempo of the French Indochina War. The impact of the seasonal monsoon pattern, endemic to the Indochina Peninsula, has been documented as a major influence in the planning and conduct of operations during the conflict (Fall 1966; Pissardy 1982; Winters et al. 1998). In addition to the tropical climate setting, perhaps one of the unique aspects of the French Indochina War is the prominence of airborne parachute drops as a means of maneuvering military forces. In reaction to the guerilla-style tactics employed by the Viet Minh, the French military relied heavily on air transportation for both resupply and combat operations (Shrader 2015). Airborne parachute drops were often conducted in conjunction with ground operations, where paratroops captured key positions immediately prior to ground assaults. With near-total air superiority, the French employed this strategy as an attempt to counter the challenges posed by the hostile and often difficult to traverse physical landscapes (Shrader 2015).

Given the prominence of airborne operations and the consequent dependence on flyable weather conditions in a wet, monsoonal climate, debate exists over the roles that weather and climate conditions played in dictating the outcome of the conflict. Some historians argue that the difficulties posed by the local physical and climatic landscapes, such as heavy rain, persistent cloud cover, and low visibility, critically limited the French forces’ ability to execute successfully airborne operations (e.g., Porch 2010). Others assert that other factors weighed more heavily in the French defeat, such as equipment and troop availability, and that the greatest impact weather conditions posed was on equipment maintenance, from rotting parachutes and rusting aircraft amid persistently humid conditions, rather than the weather conditions themselves restricting the ability to fly (e.g., Croizat 1967). Despite this ongoing debate, there has been no attempt in the historical analysis of the conflict to catalog comprehensive weather conditions during airborne operations throughout the French Indochina War. In an effort to add newly summarized evidence relevant to this debate and to demonstrate a new methodology for spatial data enrichment, this study will link weather conditions with the digital event gazetteer of French airborne operations, referred to hereafter as opérations aéroportées (OAPs; Cromley 2016). This novel method of integrating weather data into a digital event gazetteer will demonstrate how this expanded, descriptive spatial dataset can provide a more holistic view of the conflict.

c. Incorporating weather data into a digital event gazetteer

To reference the weather conditions during each OAP from the digital event gazetteer, period observational data of surface and/or upper-atmospheric variables would likely provide the most accurate representations. The most readily accessible version of such information is contained in monthly summary reports featuring a limited amount of observations for cities and airfields across Indochina, observed during the conflict. These field reports were compiled by the French Service Météorologique (NOAA 2013) and include daily surface observations, such as temperature and precipitation, as well as brief, monthly narratives reviewing some key weather events. Since the reports have ample missing data and the earliest known report during the timeframe of the French Indochina War was compiled in 1949, the reports only represent a portion of the entire conflict.

Reanalysis datasets of atmospheric variables represent a comprehensive alternative for associating weather data to OAPs, as opposed to the discontinuous observational record in Indochina during the midtwentieth century. Reanalysis datasets are model reconstructions of several weather variables across a latitude/longitude grid over time (e.g., Dee et al. 2016). In addition to serving as comprehensive data sources of weather variables free from data gaps, some recent reanalysis projects have extended further back in time to include years throughout the twentieth century and in some cases even earlier (e.g., Compo et al. 2011; Poli et al. 2016). Such gridded data sources offer the opportunity to recreate approximate daily weather characteristics for regions where observational data are sparse or unavailable, such as Indochina in the 1940s and 1950s.

Another advantage of gridded reanalysis data is that such data are ideally suited for constructing climatologies of common weather patterns experienced across a particular region over time (e.g., Huth et al. 2008). These weather patterns can be used to represent the daily and seasonal variations in atmospheric conditions, which can then be linked to observational data, as characterizations of surface conditions, whenever and wherever available. High-quality observational data became more widespread throughout the region in the years following the French Indochina War. Therefore, it is possible to use reanalysis data as a bridge to reconstruct common daily weather patterns occurring during the conflict as well as the subsequent years and then to relate modern observational data to the generalized weather patterns that occurred during the conflict, when observational data are sparse. This relationship provides an approximation of the surface conditions, such as the amount of precipitation experienced when a specific weather pattern occurs on a particular day. The weather patterns and observational data can then be codified and entered into the digital event gazetteer as additional attributes to enrich the spatial database with descriptive weather data.

d. Synoptic classification and self-organizing maps

Synoptic climatology broadly focuses on examining the controlling influence of atmospheric circulation to surface-based environmental phenomena (Davis and Kalkstein 1990; Yarnal 1993). Synoptic climatological studies often seek to address a specific set of environmental applications to larger-scale atmospheric behavior or vice versa. This is typically investigated by employing a classification scheme, which characterizes the wide variety of atmospheric conditions for a certain location over time into a manageable sample of representative weather types or patterns (Davis and Kalkstein 1990; Yarnal 1993).

Many different strategies exist for producing classifications, but one that has emerged recently in climate literature, which is also ideally suited for connecting weather patterns to discrete historical events, is the self-organizing maps (SOMs) procedure. SOMs, first detailed by Kohonen (1995) and subsequently adapted for use in synoptic climatology (e.g., Hewitson and Crane 2002; Sheridan and Lee 2011), have been used for various climate applications relating circulation patterns to environmental responses. Such applications include extreme wintertime precipitation in southeastern Europe (Cavazos 2000), investigating soil moisture interactions with atmospheric convection in the United States (Ford et al. 2015), and parsing the phases of the East Asian summer monsoon (Chu et al. 2012). One key advantage to the SOM procedure is that the resulting set of representative weather patterns are produced in a continuum, such that each pattern typically transitions into a neighboring pattern over time. This enables a unique analysis of the transition in weather patterns from day to day, week to week, and even season to season.

For this study, a synoptic classification, utilizing the SOM procedure, is used to characterize the weather conditions on all days during the French Indochina War, including those on every OAP day featured in the digital event gazetteer. Precipitation data are then associated with each of the resulting weather patterns by using the SOM procedure to classify the days during the conflict (1940s–50s) along with the days of modern, high-quality observational data (1981–2010). Then, by matching the daily weather patterns to daily precipitation observations during the modern era, the patterns can then be used to approximate wet and dry conditions during the comparatively data-sparse years of the French Indochina War, as a mechanism for assimilating descriptive weather data into a digital event gazetteer.

e. Utilizing digital event gazetteers for historical analysis

After the SOM classification creates characterizations of weather conditions that can be entered as attributes into the digital event gazetteer, the resulting enriched spatial dataset can then be applied to research questions for critical historical analysis of the French Indochina War. A proof of this concept is demonstrated by proposing a few examples of research directions that may apply within the context of the conflict. Mostern and Johnson (2008) suggest a three-prong approach to analyzing events and their attributes featured in an event gazetteer: temporal geometry, spatiotemporal setting, and causality. Events included in a historical gazetteer usually are sorted in chronological order, from the earliest to latest in time (temporal geometry). Events can then be examined to assess which events have similar themes or settings that are independent of the chronological sequence (spatiotemporal setting). Events in such a gazetteer also can be linked to establish cause–effect relationships (causality), which aid in establishing the broader narrative of the events.

The digital event gazetteer of OAPs used a temporal geometry in its base structure. The SOM weather pattern on each drop day serves as a separate attribute for each of the drops in their chronological order. This enables a comprehensive assessment of the weather patterns that coincided with each parachute drop. In this study, the weather patterns occurring on the different parachute drop days are evaluated by examining whether airborne operations tended to occur more frequently during certain weather patterns rather than others.

Other attributes for each parachute drop include the location and political region in which the drop occurred, the number of troops jumping from aircraft in each drop operation, and the type of mission the drop was fulfilling. Drops can be grouped based on these spatiotemporal or thematic settings and may exhibit certain patterns, possibly to infer relationships between weather conditions and parachute drops. As an example in this study, weather data from the gazetteer are summarized by month by political region to demonstrate an investigation into whether OAPs occurred during preferred drop seasons and when drop activity may have tended to occur during certain months in different regions of Indochina, possibly in response to seasonal weather conditions.

Additionally, historical accounts of the war include anecdotal reference to weather interfering with planned drop operations, such as prohibiting or delaying a drop (e.g., Fall 1966; Pissardy 1982). On the scale of specific airborne operations, the day-to-day and week-to-week transitions in weather patterns featured in the digital event gazetteer are briefly assessed to determine if changes in prevailing weather conditions noted in historical synopses are evident in the weather data cataloged in the digital event gazetteer. Further, such analyses can be broadened to examine if windows of preferable weather conditions emerged that seemingly enabled or contributed to parachute operations being performed. This causality-based analysis can be extended to the entire conflict as a means of assessing the flyability of different weather conditions by region over time.

2. Data and methods

a. Digital event gazetteer

The digital event gazetteer database used in this study is a spreadsheet containing 230 unique entries, each representing a parachute drop operation (OAP). These OAPs contain specific temporal and spatial data in the form of geographic coordinates (latitude and longitude) and the day, month, and year on which each of the events occurred. Each event also includes several descriptive attributes to provide greater context for each event, such as the size of the force that was dropped and the name of the specific units involved. Certain operations had code names given to them such as “Lea” or “Ceinture,” which represents a planned mission. Additionally, these OAPs are categorized by size of the element (i.e., platoon, company, battalion) and a code for one of four operation types (commando operation, reinforcement of outpost or unit under attack, independent operations, or large-scale vertical envelopment) that were created for each airborne operation.

Table 1 depicts a sample of OAP entries and associated attribute data from the digital event gazetteer. Each OAP is assigned to a location, oftentimes linked to the objective of the airborne operation. Descriptive attributes related to each OAP are provided, along with weather data produced by the reanalysis and observational data utilized in this study. Other relevant information that is currently incomplete or unavailable, such as the time of day for each operation and other weather observations, such as wind and visibility, can be included in the digital event gazetteer as continued research into these historical events uncovers new information.

Table 1.

Sample entries from the digital event gazetteer produced in Cromley (2016), enriched with weather data, including a weather pattern (SOM pattern, as shown in Fig. 3), theater and regional precipitation averages, and field observation data, where available. Additional descriptive attributes for each entry, such as the time of operation and other weather conditions can be added to the gazetteer as continuing archival research uncovers new information.

Table 1.

b. Synoptic classification

In general, weather conditions must exceed certain established flight minimums to enable parachute operations, unless critical situations require dropping in less ideal conditions (Pissardy 1982). Weather conditions such as visibility, cloud ceiling, and low-level and surface wind speeds are critical for successfully executing parachute drop operations [Departments of the Army and the Air Force (DAAF) 1967]. These weather parameters, particularly wind speeds, can vary considerably throughout the day in tropical climates. Given the general lack of available weather observations during drop operations, and that the time of day (hour) for each OAP is largely unknown, performing a synoptic classification using mean sea level pressure (MSLP) can serve to characterize weather conditions. Classifications can be performed using other reanalysis variables, but MSLP is a robust and widely used reanalysis variable and is convenient as a single variable that can broadly represent a range of weather conditions important for OAPs, including cloud cover and precipitation.

To classify weather patterns in the context of the French Indochina War, MSLP data were acquired from the European Centre for Medium-Range Weather Forecasts (ECMWF) Twentieth Century Reanalysis (ERA-20C) dataset (ECMWF 2014; Poli et al. 2016) at a resolution of 1° latitude × 1° longitude. The ERA-20C reanalysis represents a recent effort to reconstruct weather conditions throughout the twentieth century that interpolates weather variables based in part on observations, including surface pressure (Poli et al. 2016). The number of surface pressure observations included in the data assimilation increase throughout the twentieth century, so data accuracy generally improves through time. During the 1940s–50s, the era of the French Indochina War, ERA-20C reanalysis data have been shown to be reasonably accurate, with surface pressure errors in the tropics generally less than 5 hPa, with even lower error as time progresses (Poli et al. 2013).

Daily 0000 UTC MSLP values were collected from 1 January 1945 to 31 December 2010 across a domain from 8° to 24°N and from 95° to 118°E, an area of study that encompasses the Indochina Peninsula as well as surrounding portions of the South China Sea to the east, Gulf of Thailand to the south, and Andaman Sea to the west (Fig. 1). Because of a lack of comprehensive and readily accessible surface or upper air weather observations in the region spanning the French Indochina War, the SOM classification of MSLP is performed across all days from 1945 to 2010. This was done to inform the resulting classifications based on weather patterns both during the conflict as well as during a more modern era, for which comprehensive, high-quality weather observations for the region are available. Broadly speaking, by comparing modern precipitation data to the classification, the patterns that correspond to relatively high or low precipitation in the modern era also likely represent comparatively wet and dry patterns during the years of the French Indochina War.

Fig. 1.
Fig. 1.

Map of the study area, including the selected reanalysis domain (dashed line), selected GHCN stations coded by data reporting percentage for the modern era (1981–2010), and the locations of all OAPs included in the Cromley (2016) gazetteer.

Citation: Weather, Climate, and Society 10, 1; 10.1175/WCAS-D-17-0016.1

To construct a synoptic classification of the climatology of Indochina using the SOM technique, a principal component analysis (PCA) was first performed for the daily MSLP data to reduce autocorrelation between reanalysis grid points. Seven PCs were retained, each with eigenvalues of at least 1.0, explaining 98.6% of the total variance (Guttman 1954; Kaiser 1960; Yarnal 1993). The SOM procedure was then run using these retained principal components as inputs. The cluster centroids (referred to as nodes) resulting from the SOM procedure arrange in a latticelike structure, such that the nodes located near each other in data space are more similar than nodes located farther apart. This arrangement carries over to the resulting maps of synoptic classifications, with similar patterns grouped together. Consequently, the days assigned to each node/pattern tend to transition into neighboring nodes/patterns, which feature similar, but slightly different, weather patterns, a trend that generally echoes the transitional nature of the atmosphere.

As in most applications of clustering, the selected number of clusters has important implications. The user can define the dimensions of the SOM lattice based on multiple considerations. These include the representativeness of the topology of the fitted node lattice to the input data distribution; the distribution of the number of data points assigned to the resultant nodes; the balance between the minimization of within-cluster variability and the maximization of between-cluster variability, assessable via clustering evaluation techniques (i.e., Davies and Bouldin 1979); and simply the end purpose of the clustering overall—in this case, producing a representative synoptic classification that effectively stratifies mean sea level pressure patterns (and the precipitation averages associated to them) across Indochina for the purpose of relating weather parameters to specific airborne operations during the French Indochina War cataloged in the digital event gazetteer.

Multiple SOM dimensions were examined and, based on the above considerations, an 8 × 4 lattice (32 total nodes) was selected as the ideal configuration. This configuration was found to be robust after additional testing, including alterations to various settings such as the learning rates, the total number of learning iterations, and neighborhood radius, each producing similar resultant cluster solutions and comparable Davies–Bouldin index values that suggest ideal cluster representation. Each of the 32 nodes are then used to represent a MSLP field across the area of study, so that each day from 1945 to 2010 could be assigned to one of the nodes from the SOM lattice.

c. Precipitation observations

Daily precipitation observations were collected from the Global Historical Climatology Network (GHCN) via the National Center for Environmental Information (NCEI)–National Climatic Data Center (NCDC) online map data portal (accessible online at https://gis.ncdc.noaa.gov/maps/ncei/cdo/daily) from 1981 to 2010 for all stations within the reanalysis domain (Carbon Dioxide Information Analysis Center et al. 1995). Given the highly varied topographic and climatic landscapes of the Indochina Peninsula and surrounding regions, it is preferable to represent as many different locations as possible when averaging precipitation between stations across the reanalysis domain. Many of the stations feature sporadic and discontinuous periods of record however; so, after assessing the data quality and reporting percentages of these stations, a subset of the higher-quality stations was selected, representing as much of the reanalysis domain as possible (Fig. 1).

The subset includes the stations with the highest overall percentage of days reporting (≥75% days reporting) within the area of study: 39 stations in Thailand, 19 in China, and 1 in Burma. Stations located within French Indochina (present-day Cambodia, Laos, and Vietnam), where the OAPs occurred, are comparatively sparse and of lower quality. In an effort to represent locations closer to the OAPs, all sites within these countries with ≥40% of days reporting were selected. This included all stations in Vietnam (15 stations across Tonkin, Annam, and Cochinchina), where most of the OAPs occurred, but excluded all six available stations located in Cambodia and Laos, where the highest reporting station only included 14% of days. In general, the Vietnam stations have shorter periods of record as compared to the other selected stations. Therefore, representations of surface conditions (precipitation) in this region may be less representative than in other locations of the reanalysis domain. The selected subset of 74 total stations was used to calculate theaterwide and regional-scale daily precipitation averages.

When and where available, the field weather observations taken during the conflict can also be entered as attributes in the digital event gazetteer for specific OAPs. These reports, entitled “Résumé Mensuel du Temps en Indochine,” contain daily precipitation observations, aggregated by month (NOAA 2013). Digital scans of the reports from 1949 to 1952 were made available as part of the National Oceanic and Atmospheric Administration (NOAA) Environmental Data Rescue Program and were accessed online from the NOAA Central Library (NOAA 2013). While the modern NCEI–NCDC precipitation data were used to characterize the weather conditions associated with each synoptic classification at the theater and regional scales, the field weather observations from the historical reports serve as supplementary data entered into the digital event gazetteer, which may contribute to the analysis of OAPs at the campaign to specific operation levels.

d. Methodology

The overall methodology for enriching a digital event gazetteer with weather data demonstrated in this study can be presented as a framework, summarized in Fig. 2. The digital event gazetteer catalogs historical events with spatiotemporal information. To associate descriptive weather attributes with each OAP entry in the gazetteer, the weather factors pertinent to OAPs must be identified, such as those provided in field manuals or other archival references. If available, historical weather observations provide an acceptable resource for referencing weather conditions during each OAP. With the date and location for each OAP, the important weather factors identified, and access to a resource for historical weather observations, a decision must be made as to whether the available weather observations are suitable as descriptive attributes for each OAP. In the example demonstrated in this study, while historical weather observations are available, they are limited in coverage and primarily represent larger cities, which are distant from the OAPs (~95.5 km on average).

Fig. 2.
Fig. 2.

Methodological framework for enriching a digital event gazetteer with weather data based on historical observations or reanalysis data. Note that the framework can be implemented for enriching digital gazetteers of other events using different weather variables.

Citation: Weather, Climate, and Society 10, 1; 10.1175/WCAS-D-17-0016.1

Limited weather observations from the historical logs can be entered into the gazetteer, but for a more comprehensive characterization of weather conditions for every drop, a synoptic classification using reanalysis data produces representative weather patterns, which can in turn be related to surface conditions, as measured by modern observations. These descriptive weather data are then entered for each drop cataloged in the digital event gazetteer, thereby providing an approximate characterization of weather conditions for each event. In this study, mean sea level pressure and precipitation are used as sample weather data, but this procedure can be replicated for other relevant reanalysis and observed weather variables. Similarly, the overall methodological framework can be applied to other examples of digital event gazetteers where the enrichment of weather data is desired.

3. Discussion

a. Climatology of French Indochina

The resulting weather patterns from the SOM classification are shown in Fig. 3. Broadly, the region is dominated by two seasonal pressure patterns, each reflecting the thermal characteristics of the season. Regimes of low pressure are centered over the Gulf of Tonkin and South China Sea during the warm season (patterns toward the left of the figure), and high pressure regimes are centered over mainland China during the cool season (patterns toward the right of the figure). These opposing pressure patterns represent the response to an annual monsoon, featuring a seasonal reversal of winds across the Indochina Peninsula from dominant westerlies and southwesterlies during the warm season to easterlies and northeasterlies in the cool season (Chang et al. 2005).

Fig. 3.
Fig. 3.

Daily mean sea level pressure patterns resulting from the SOM classification procedure. Darker blues indicate lower pressure, while darker reds indicate higher pressure. OAP locations (black dots) are plotted by the SOM pattern corresponding to each drop date.

Citation: Weather, Climate, and Society 10, 1; 10.1175/WCAS-D-17-0016.1

The complex monsoon patterns over Indochina and the surrounding region have been well documented in previous studies (e.g., Meehl 1987; Li and Yanai 1996; Wang and LinHo 2002; Zhang et al. 2002; Takahashi and Yasunari 2006), including literature from the era of the French Indochina War (e.g., McCune 1947; Sanderson 1954). The monsoon was a recognized feature of the climate known to the French military, as exhibited by its frequent reference, both in military synopses of the conflict (e.g., Pissardy 1982) and in the monthly weather reports included in the “Résumé Mensuel du Temps en Indochine” field observation logs (NOAA 2013).

To examine further the climatology produced by the synoptic classification, the percentages of all days by SOM pressure patterns are shown in Table 2a. The most common patterns tend to be the seasonally opposing extremes of low and high pressure centered over southern China and northern portions of Indochina. As the year progresses, patterns generally transition into neighboring patterns, such that by spring and fall when the monsoon winds are transitioning, the most common patterns tend to be those that feature weaker versions of high and low pressure (patterns located toward the center of Fig. 3), with notably less pronounced centers of circulation as compared to the seasonal extremes.

Table 2.

(a, top) Percentages of all days (1945–2010) by SOM pressure pattern, (b, middle) theaterwide precipitation averages by SOM pattern in millimeters (mm), and (c, bottom) percentages of all OAPs (parachute drops) by SOM pattern. The data in each section of the table are arranged in the same lattice as the SOM patterns shown in Fig. 3, such that pattern 1 is positioned in the lower left of each table section and pattern 32 is in the upper right. Numbers in bold are equal to or greater than the average for all patterns in each section: (top) 3.125%, (middle) 4.3 mm, and (bottom) 3.1%.

Table 2.

In assessing the precipitation associated with each SOM pattern, the theaterwide average of precipitation (average of all selected GHCN stations) from the modern observational data clearly stratifies precipitation in accordance to expectations (Table 2b). The low pressure patterns of the warm season generally feature the wettest conditions on average across the entire study region, while the high pressure patterns of the cool season are driest. Convection is commonplace across Indochina in the summertime, particularly in the south, because of southwesterly monsoon winds and daytime heating in the humid tropical climate (Pham et al. 2010), so it is unsurprising for the common summer patterns to yield higher amounts of precipitation on average across Indochina.

Conversely, the wintertime is cooler and generally drier across Indochina, particularly across the interior and toward the south (Pham et al. 2010). In Tonkin, toward the north of the peninsula, low-level clouds, termed the crachin, drape across the highlands during the cool months and were known to pose unique challenges for combat operations (Sanderson 1954; Winters et al. 1998). Since the theaterwide precipitation averages taken from station observations between 1981 and 2010 are stratified intuitively across the SOM patterns based on MSLP data from 1945 to 2010, the likelihood that the classification in this study is valid is enhanced and can be used to associate MSLP patterns with precipitation both in the modern era and during the French Indochina War.

b. Weather patterns and OAPs

The climatology produced by the synoptic classification links each day of the French Indochina War to a discrete weather pattern. Thus, the pattern associated with each OAP day can be identified and then entered as an attribute for that OAP record in the digital event gazetteer. Theaterwide precipitation averages are also summarized by synoptic pattern and added to each record in the gazetteer. These provide approximations of the potential surface weather conditions observed during each OAP of the French Indochina War.

With the digital event gazetteer enriched with descriptive weather data, historical analysis into potential relationships between OAPs and weather conditions is possible. For example, the percentages of all OAPs coinciding with each weather pattern can be summarized, as shown in Table 2c. OAPs occurred at least once during every SOM pattern, resulting from the classification and tended to occur most frequently during patterns that dominate the transition months, particularly from March to May and in October. This is reflected in the higher percentages in Table 2a for patterns depicting the less amplified low/high pressure in the transition seasons.

The theaterwide precipitation averages convey a similar connection to OAPs, with moderate precipitation totals generally observed during these transition patterns (Table 2b). The summertime low pressure regimes are generally the wettest on average (~6.0–8.0 mm) and the wintertime high pressure regimes among the lowest (~1.0–3.0 mm). The OAPs occurred most frequently during the transitional weather patterns, when theaterwide precipitation averages ~3.0–6.0 mm, patterns that dominate the spring and fall months.

Sanderson (1954, p. 69) remarked on the challenges that the climatic seasons posed toward military action in Indochina, determining that while ground operations are “hampered the year around by poor soil trafficability,” conditions are made even worse during the wet summer months, rendering roads “almost useless when pounded by rain and rutted by vehicular traffic.” Airborne operations are also limited during the summer, mainly due to low cloud cover, though “some air support can be rendered even in summer, if the techniques of extremely low flight are used.” The winter by contrast “offers the best opportunity for air support operations,” due to drier conditions, particularly over the interior. Sanderson concludes that while “over the balance of the country [Indochina] conditions in winter are at their best for large-scale military campaigns by a mechanized army…the summer season offers little hope for large-scale, modern warfare.” The average precipitation linked to the seasonal pressure patterns corroborates these general observations of the regional climate, such that these descriptive weather attributes may serve to inform historical analyses of OAPs.

While the weather data produced and compiled in this study can be used to assess day-to-day weather transitions, it should be noted that shorter-term and smaller-scale weather variations, such as day-to-day changes, are not as clearly represented as longer-term and larger-scale transitions, such as by week, month, or season. For example, individual storm events, such as tropical cyclones, are not as well represented by the resulting synoptic classification patterns as longer-term variations, such as the seasonal monsoon. Similarly, precipitation observations featured for specific stations among the “Résumé Mensuel du Temps en Indochine” logs can vary substantially from the theater and regional precipitation averages produced for the study area, particularly because of the localized nature of common tropical weather phenomena, such as thunderstorms. Historical analyses using weather data–enriched gazetteers should consider the spatiotemporal scale and the representativeness of featured attributes.

c. Regional precipitation and OAPs

Although theaterwide precipitation is stratified across the SOM patterns in an intuitive manner, it should be noted that the locations across the Indochina Peninsula experience the wet and dry seasons differently. The peninsula features complex topography, including the Annamite Range, a roughly north–south-oriented ridgeline reaching ~2500–3000 m in elevation, spanning much of the present-day border of Vietnam and Laos (Abramov et al. 2006). Because of this, as well as the influences of the East and South Asian monsoon forcings (Zhang et al. 2002), each region of the Indochina Peninsula experiences wet seasons during different times of the year, in response to the seasonal reversal of prevailing wind direction relative to the terrain (Chen et al. 2012; Wu and Hsu 2016).

Such seasonal weather idiosyncrasies can be summarized by region using the digital event gazetteer. Synopses of the French Indochina War often use political boundaries to assess the various campaigns that took place throughout the conflict (e.g., Pissardy 1982); thus, the political region associated with each OAP is included in the gazetteer. To characterize weather conditions by region, the modern precipitation observations were aggregated by political boundary for the three regions featuring the most OAPs (Tonkin, Annam, and Cochinchina) to construct regional precipitation averages as additional gazetteer attributes. Figure 4 depicts the average daily precipitation by month for these three regions. The numbers of OAPs in each region are also aggregated by month to assess the relationships between relatively wet and dry months to the peak months of parachute drops.

Fig. 4.
Fig. 4.

Graphs of average daily precipitation for each region (solid line) and theaterwide (dashed line) aggregated by month by region. The total number of OAPs during the conflict per month in that region is shown in bars. No OAPs occurred in Annam in November or December.

Citation: Weather, Climate, and Society 10, 1; 10.1175/WCAS-D-17-0016.1

As shown in Fig. 4, generally, across each region, an inverse relationship exists between the frequency of OAPs and the peak months for average daily regional precipitation. Correlations between monthly drop frequency and precipitation range from −0.31 to −0.53 across the regions, with each result significantly different from zero at α = 0.05. In Tonkin, for example, the wettest months in the region (June–August) also featured the fewest OAPs. The peak months for parachute drops occurred in October and in April–May, when the average daily precipitation is not at its peak, though there are drier months on average, so the inverse relationship is not perfect. This suggests the importance of other factors critical to planning drop operations, such as enemy activity, equipment/personnel availability, and the type and scale of the operation. While it is likely that weather affected the occurrence of OAPs, it is not the only factor, as these other considerations are also strong contributors.

The general pattern also holds true for Annam and Cochinchina, where the peak months for OAPs occur during slightly drier months, as compared to the peak precipitation months. Notably, the wettest months occur in the fall for Annam, when the OAPs are least common month to month. Cochinchina features a similar pattern and precipitation–parachute drop relationship to Tonkin, with a slightly longer-duration wet season from June to October. The precipitation trends observed in each region and the geographic distribution of differing wet-season definitions generally align with previous findings and characterizations (i.e., Pham et al. 2010; Chen et al. 2012), further suggesting that, despite some known variability, the wet and dry seasons recur annually and consequently they likely bore some influence on military activity during the French Indochina War. This apparent connection between OAP frequency and precipitation is uniquely discernable from the information cataloged in the digital event gazetteer and warrants further investigation. While political boundaries serve as convenient spatiotemporal settings for aggregation, given the topography of Indochina, averaging precipitation by other spatiotemporal themes and boundaries, such as by climate regions, may produce different results that could then facilitate further investigation into these potential relationships.

d. Individual operation analysis

The structure of the weather data–enriched digital event gazetteer also enables historical research into possible cause–effect relationships between weather conditions either enabling or preventing OAPs from occurring. To demonstrate this, a case study is briefly examined of the Hoa Binh campaign, which featured a relatively regular series of parachute drops conducted in the vicinity of Hoa Binh in Tonkin from October 1951 to January 1952. The campaign began in reaction to Viet Minh movements, with the first three OAPs taking place in rapid succession (2, 4, and 6 October). As the campaign evolved, the French eventually planned their own activity, orchestrating larger drop operations. Subsequent OAPs occurred in support of the ongoing mission, in part driven by earlier drops.

As the dominant sea level pressure patterns change over the seasons in Tonkin, both in response to temperature and wind direction, this progression can be seen in the average precipitation by pattern, plotted by day. Figure 5 and Table 3 depict the evolution of weather patterns in Tonkin during the time of the Hoa Binh campaign. The average wet season in Tonkin is broadly defined as May–September, so as October progresses, the patterns gradually begin to transition into drier conditions as the dry monsoon emerges. Of note, however, variations in day-to-day weather conditions are evident as the SOM pattern (represented in Fig. 5 by the corresponding precipitation) changes between neighboring patterns.

Fig. 5.
Fig. 5.

Daily regional precipitation averages in Tonkin for all days associated with each day’s assigned SOM pattern for September 1951–February 1952. The inset graph depicts the same data for all days in 1951–52, with September–February highlighted. Stars indicate days on which one or more OAPs occurred during the Hoa Binh campaign in Tonkin.

Citation: Weather, Climate, and Society 10, 1; 10.1175/WCAS-D-17-0016.1

Table 3.

Calendar of selected days before and after OAPs during the Hoa Binh campaign. The SOM pattern assigned to each day and the regionally averaged precipitation in Tonkin for all days associated with each listed day’s SOM pattern. Rows in bold indicate dates on which at least one OAP occurred.

Table 3.

The first OAP of the Hoa Binh campaign occurred on 2 October on a day represented with a relatively dry pattern (SOM pattern 8, with weak high pressure over the region and 1.2 mm of precipitation on average in Tonkin), as shown in Table 3. The two subsequent OAPs occur on 4 and 6 October, each on days with slightly wetter patterns (SOM patterns 4 and 5, weak low pressure in the region; 4.3 and 3.5 mm of precipitation, respectively). These operations all occur as a reaction to Viet Minh aggressions in the region, so it is possible that the evolution of the war necessitated these drops, regardless of antecedent weather and ground conditions. These relatively medium-sized OAPs featured 500–700 troops and were considered commando and reinforcement mission types. Considering the challenges discussed by Sanderson (1954), it is possible that smaller parachute operations could be performed during periods of less ideal weather conditions, particularly for high priority commando and reinforcement OAPs, as during such operations, troops can be injected directly into hostile areas without needing to wade through poor surface conditions en route, as occurred during the siege of Dien Bien Phu in 1954 (Fall 1966). Smaller drops also require fewer aircraft and take less time to complete, suggesting that drops could occur during poor conditions or breaks in inclement weather. For example, on 13 October 1949, 131 paratroopers dropped in “terrible weather, under heavy rain and violent wind” (Pissardy 1982, p. 94). On 7 December 1948, 408 paratroopers dropped when “…weather conditions are bad. Between noon and 2:30 pm, a thinning allows the release [of troops]…a further deterioration of the weather precludes the ‘dropping’ of [more troops] in the afternoon” (Pissardy 1982, p. 84).

By contrast, larger, planned offensive operations would likely prioritize favorable weather conditions, particularly if heavy equipment and large numbers of ground forces are needed to assemble in a particular area or if thousands of troops are planned to jump, requiring dozens of aircraft flying in formation. For example, prior to Operation Castor, the initial French assault to capture Dien Bien Phu in November 1953, French commanders and weather personnel flew a reconnaissance mission to determine if conditions were favorable for the airborne operation. Fog blanketed the drop zone on the morning of the first drops, but the clouds soon cleared, thereby enabling the largest troop transport operation of the entire conflict (Fall 1966). In general, such major operations would likely be planned to occur during a period of persistently drier conditions to allow ground conditions to improve from the summer rains, to increase the likelihood of flyable weather, and to maximize the use of the dry season for operations (Winters et al. 1998). These considerations may be evident in the inverse relationships noted between OAPs and monthly precipitation aggregated by political region, suggesting a preference for performing OAPs during the drier months.

As time progressed into November and December of 1951, the course of the Hoa Binh campaign enabled the French to plan larger drops during these months as they went on the offensive in the region. OAPs in this timeframe, particularly in December, occurred on days featuring drier patterns. These planned drops required extensive ground movements, so it is likely that the French considered the weather in their planning for these operations, as was noted for Operation Castor (Fall 1966). Extended periods of drier patterns emerge in December, around the time of a few OAPs, with a few slightly wetter patterns embedded in between. From 14 November to 12 February, no pattern that occurred averages more than 3.0 mm of precipitation, which suggests a lengthy window of generally favorable conditions for these operations to occur. Five of the OAPs during the campaign occurred during this window. Wetter patterns became more common in February and beyond, after the drops had been completed, so the Hoa Binh campaign season in Tonkin seemingly coincided with the dry season in Tonkin, a period of relatively favorable ground and atmospheric conditions.

To supplement the analysis of weather and OAPs on individual drop days, field weather observations from the “Résumé Mensuel du Temps en Indochine” logs are included in the digital event gazetteer when available, which include the drops days during the Hoa Binh campaign. The observations generally align with the indications of the regional precipitation averages. For the first three reaction OAPs, the closest field precipitation observations were 10.0, 41.3, and 32.9 mm, respectively. The observation for the first OAP comes from Haiphong, a station ~283 km away, and the other two come from Hanoi, ~182 km away. The heavier precipitation recorded at Hanoi during the second and third OAPs coincides with the day-to-day transition to wetter SOM patterns on those days in particular. The November/December planned OAPs occur on days with an average of 0.4 mm observed at the nearest reporting stations, generally 300–400 km away. Accordingly, the SOM patterns transition to drier patterns during these days, so the SOM classification seemingly recognizes these day-to-day transitions in weather. The field observations further suggest that the SOM patterns and the associated regional precipitation averages are reasonably representative of conditions during these days in the war. Therefore, the weather data–enriched digital event gazetteer produced in this study can be utilized for further study of individual operations.

4. Summary and conclusions

Digital event gazetteers are datasets of spatiotemporal information regarding a series of historical events. Adding weather data to the gazetteer enables new perspectives regarding the relationships between space, time, and the events. This study outlines a methodology for incorporating weather characteristics for each event included in a gazetteer, for times and locations where weather observations are largely unavailable. Through a synoptic classification using reanalysis data, representations of weather patterns were created for a dataset of parachute operations during the French Indochina War. Modern precipitation observations were associated with each weather pattern, as an approximation of the surface conditions of each weather pattern, including those during the conflict. With the addition of this weather information into the gazetteer, the relationships between parachute operations and their location in time and space can be analyzed in greater detail, including how monsoonal seasons may have affected parachute operations. Consequently, new research questions can emerge about these events, potentially adding new evidence to an ongoing historical debate over the role the French military dependency on parachute operations in Indochina played in determining the outcome of the conflict and if, or to what extent, the weather and regional climates may have been contributing factors.

While observations from the era provide the best available characterizations of the weather during each of the parachute operations, such observations are limited in coverage. A synoptic classification of reanalysis data presents a comprehensive alternative that, while subject to its own limitations (e.g., Bengtsson et al. 2004), can yield valid and meaningful results. The methodology described here can be applied to additional weather variables, such as temperature, wind, and humidity, to provide further evidence relative to the debate over the role of weather during the French Indochina War as well as to other datasets of discrete events in general where the inclusion of weather data can be of use from geophysical phenomena to cultural events. The example analyses outlined in this study also show how weather data–enriched event gazetteers can inform additional lines of inquiry, which were previously not possible without the assimilation of weather and event data.

With the ever-expanding availability and collection of spatiotemporal data, it is essential to organize such data into efficient, manageable, and functional datasets to facilitate analysis. Further, it is important also to consider the applications of data as justifications for its collection, as information can be applied in different ways to unveil relationships that could otherwise go unnoticed if the data were not combined. As digital event gazetteers and synoptic classifications continue to evolve over time, it is likewise important to recognize how combining these and other datasets can offer compatible data to improve the understanding of historical events.

Acknowledgments

The authors wish to acknowledge Cameron Lee and Scott Sheridan for sharing their expertise and offering assistance in developing the synoptic classification produced in this study, Col. Andrew Lohman for providing background information and resources regarding airborne operations, and Robert G. Cromley and Thomas Schmidlin for offering feedback on the manuscript. Both authors acknowledge the Department of Geography at Kent State University for their support of this research, including providing funding in support of conference presentations regarding this work. The authors also thank three anonymous reviewers whose insightful comments improved this manuscript.

REFERENCES

  • Abramov, A. V., V. V. Rozhnov, and P. N. Morozov, 2006: Notes on mammals of the Ngoc Linh Nature Reserve (Vietnam, Kon Tum Province). Russ. J. Theriology, 2, 8592.

    • Search Google Scholar
    • Export Citation
  • Bengtsson, L., S. Hagemann, and K. I. Hodges, 2004: Can climate trends be calculated from reanalysis data? J. Geophys. Res., 109, D11111, doi:10.1029/2004JD004536.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bodenhamer, D. J., J. Corrigan, and T. M. Harris, 2010: The Spatial Humanities: GIS and the Future of Humanities Scholarship. Indiana University Press, 222 pp.

  • Buckland, M., A. Chen, F. C. Gey, R. R. Larson, R. Mostern, and V. Petras, 2007: Geographic search: Catalogs, gazetteers, and maps. Coll. Res. Libr., 68, 376387, doi:10.5860/crl.68.5.376.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Carbon Dioxide Information Analysis Center, Environmental Sciences Division, Oak Ridge National Laboratory, U. S. Department of Energy, Arizona State Climate Office/Arizona State University, and National Climatic Data Center/NESDIS/NOAA/U.S. Department of Commerce, 1995: Global Historical Climatology Net (GHCN) version 2 temperature, precipitation, pressure. Research Data Archive at the National Center for Atmospheric Research, Computational and Information Systems Laboratory, accessed 15 June 2017, http://rda.ucar.edu/datasets/ds564.0/.

  • Cavazos, T., 2000: Using self-organizing maps to investigate extreme climate events: An application to wintertime precipitation in the Balkans. J. Climate, 13, 17181732, doi:10.1175/1520-0442(2000)013<1718:USOMTI>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chang, C. P., Z. Wang, J. McBride, and C. H. Liu, 2005: Annual cycle of Southeast Asia—Maritime continent rainfall and the asymmetric monsoon transition. J. Climate, 18, 287301, doi:10.1175/JCLI-3257.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chen, T. C., J. D. Tsay, M. C. Yen, and J. Matsumoto, 2012: Interannual variation of the late fall rainfall in central Vietnam. J. Climate, 25, 392413, doi:10.1175/JCLI-D-11-00068.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chu, J. E., S. N. Hameed, and K. J. Ha, 2012: Nonlinear, intraseasonal phases of the East Asian summer monsoon: Extraction and analysis using self-organizing maps. J. Climate, 25, 69756988, doi:10.1175/JCLI-D-11-00512.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Compo, G. P., and Coauthors, 2011: The Twentieth Century Reanalysis Project. Quart. J. Roy. Meteor. Soc., 137, 128, doi:10.1002/qj.776.

  • Croizat, V. J., 1967: A translation from the French: Lessons of the war in Indochina. Vol. 2. Rand Corporation Memo. RM-5271-PR., 415 pp, http://www.rand.org/pubs/research_memoranda/RM5271.html.

  • Cromley, G., 2016: Designing a military event gazetteer: The case of parachute operations during the French Indochina War. Prof. Geogr., 68, 249260, doi:10.1080/00330124.2015.1062702.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • DAAF, 1967: U.S. Army/U.S. Air Force doctrine for airborne operations. Field Manual 57-1, Air Force Manual 2-51, 71 pp.

  • Davies, D. L., and D. W. Bouldin, 1979: A cluster separation measure. IEEE Trans. Pattern Anal. Mach. Intell., PAMI-1, 224227, doi:10.1109/TPAMI.1979.4766909.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davis, R. E., and L. S. Kalkstein, 1990: Development of an automated spatial synoptic climatological classification. Int. J. Climatol., 10, 769794, doi:10.1002/joc.3370100802.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dee, D., and Coauthors, Eds., 2016: The climate data guide: Atmospheric reanalysis: Overview & comparison tables. Accessed 15 June 2017, https://climatedataguide.ucar.edu/climate-data/atmospheric-reanalysis-overview-comparison-tables.

  • ECMWF, 2014: ERA-20C Project (ECMWF Atmospheric Reanalysis of the 20th Century). Research Data Archive at the National Center for Atmospheric Research, Computational and Information Systems Laboratory, accessed 15 June 2017, doi:10.5065/D6VQ30QG.

    • Crossref
    • Export Citation
  • Fall, B. B., 1966: Hell in a Very Small Place: The Siege of Dien Bien Phu. Da Capo Press, 568 pp.

  • Fall, B. B., 1994: Street Without Joy: The French Debacle in Indochina. Stackpole Books, 416 pp.

  • Ford, T. W., S. M. Quiring, O. W. Frauenfeld, and A. S. Rapp, 2015: Synoptic conditions related to soil moisture-atmosphere interactions and unorganized convection in Oklahoma. J. Geophys. Res. Atmos., 120, 11 51911 535, doi:10.1002/2015JD023975.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Goodchild, M. F., and L. L. Hill, 2008: Introduction to digital gazetteer research. Int. J. Geogr. Inf. Sci., 22, 10391044, doi:10.1080/13658810701850497.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Guttman, L., 1954: Some necessary conditions for common-factor analysis. Pyschometrika, 19, 149161, doi:10.1007/BF02289162.

  • Hewitson, B. C., and R. G. Crane, 2002: Self-organizing maps: Applications to synoptic climatology. Climate Res., 22, 1326, doi:10.3354/cr022013.

  • Hill, L. L., 2000: Core elements of digital gazetteers: Placenames, categories, and footprints. Research and Advanced Technology for Digital Libraries: Proc. Fourth European Conf., Lisbon, Portugal, ECDL, 280–290, doi:10.1007/3-540-45268-0_26.

    • Crossref
    • Export Citation
  • Huth, R., C. Beck, A. Philipp, M. Demuzere, Z. Ustrnul, M. Cahynová, J. Kyselý, and O. E. Tveito, 2008: Classifications of atmospheric circulation patterns: Recent advances and applications. Ann. N. Y. Acad. Sci., 1146, 105152, doi:10.1196/annals.1446.019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kaiser, H. F., 1960: The application of electronic computers to factor analysis. Educ. Psychol. Meas., 20, 141151, doi:10.1177/001316446002000116.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Keller, T., 2009: The mountains roar: The Alps during the Great War. Environ. Hist., 14, 253274, doi:10.1093/envhis/14.2.253.

  • Kohonen, T., 1995: Self-Organizing Maps. Springer-Verlag, 362 pp.

    • Crossref
    • Export Citation
  • Li, C., and M. Yanai, 1996: The onset and interannual variability of the Asian summer monsoon in relation to land–sea thermal contrast. J. Climate, 9, 358375, doi:10.1175/1520-0442(1996)009<0358:TOAIVO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Logevall, F., 2012: Embers of War: The Fall of an Empire and the Making of America’s Vietnam. Random House, 864 pp.

  • McCune, S., 1947: The diversity of Indochina’s physical geography. Far East. Quart., 6, 334344, doi:10.2307/2049430.

  • Meehl, G. A., 1987: The annual cycle and interannual variability in the tropical Pacific and Indian Ocean regions. Mon. Wea. Rev., 115, 2750, doi:10.1175/1520-0493(1987)115<0027:TACAIV>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mostern, R., 2011: “Dividing the Realm in Order to Govern”: The Spatial Organization of the Song State. Harvard University Press, 396 pp.

    • Crossref
    • Export Citation
  • Mostern, R., and I. Johnson, 2008: From named place to naming event: Creating gazetteers for history. Int. J. Geogr. Inf. Sci., 22, 10911108, doi:10.1080/13658810701851438.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • NOAA, 2013: French Indochina climatological data: Résumé mensuel des observations. Accessed 11 January 2017, https://www.lib.noaa.gov/collections/foreign_climate_data_pages/foreign_climate_data_indochina.html.

  • Pham, X. T., B. Fontaine, and N. Philippon, 2010: Onset of the summer monsoon over the southern Vietnam and its predictability. Theor. Appl. Climatol., 99, 105113, doi:10.1007/s00704-009-0115-z.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pissardy, J. P., 1982: Paras d’Indochine: 1944-1954. Société de Production Littéraire, 225 pp.

  • Poli, P., and Coauthors, 2013: The data assimilation system and initial performance evaluation of the ECMWF pilot reanalysis of the 20th-century assimilating surface observations only (ERA-20C). ERA Rep. Series 14, 62 pp., https://www.ecmwf.int/sites/default/files/elibrary/2013/11699-data-assimilation-system-and-initial-performance-evaluation-ecmwf-pilot-reanalysis-20th.pdf.

  • Poli, P., and Coauthors, 2016: ERA-20C: An atmospheric reanalysis of the twentieth century. J. Climate, 29, 40834097, doi:10.1175/JCLI-D-15-0556.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Porch, D., 2010: The French Foreign Legion: A Complete History of the Legendary Fighting Force. HarperCollins, 728 pp.

  • Sanderson, R. W., 1954: Notes on the climate of Indochina. Weatherwise, 7, 5669, doi:10.1080/00431672.1954.9930321.

  • Sheridan, S. C., and C. C. Lee, 2011: The self-organizing map in synoptic climatological research. Prog. Phys. Geogr., 35, 109119, doi:10.1177/0309133310397582.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shrader, C. R., 2015: A War of Logistics: Parachutes and Porters in Indochina, 1945–1954. University of Kentucky Press, 488 pp.

    • Crossref
    • Export Citation
  • Southall, H., R. Mostern, and M. L. Berman, 2011: On historical gazetteers. Int. J. Hum. Arts Comput., 5, 127145, doi:10.3366/ijhac.2011.0028.

    • Search Google Scholar
    • Export Citation
  • Takahashi, H. G., and T. Yasunari, 2006: A climatological monsoon break in rainfall over Indochina—A singularity in the seasonal march of the Asian summer monsoon. J. Climate, 19, 15451556, doi:10.1175/JCLI3724.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, B., and LinHo, 2002: Rainy season of the Asian–Pacific summer monsoon. J. Climate, 15, 386398, doi:10.1175/1520-0442(2002)015<0386:RSOTAP>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Winters, H. A., G. E. Galloway Jr., W. J. Reynolds, and D. W. Rhyne, 1998: Battling the Elements: Weather and Terrain in the Conduct of War. The Johns Hopkins University Press, 336 pp.

  • Wu, C. H., and H. H. Hsu, 2016: Role of the Indochina Peninsula narrow mountains in modulating the East Asian–western North Pacific summer monsoon. J. Climate, 29, 44454459, doi:10.1175/JCLI-D-15-0594.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yarnal, B., 1993: Synoptic Climatology in Environmental Analysis: A Primer. Belhaven Press, 256 pp.

  • Zhang, H., T. Li, B. Wang, and G. Wu, 2002: Onset of the summer monsoon over the Indochina Peninsula: Climatology and interannual variations. J. Climate, 15, 32063221, doi:10.1175/1520-0442(2002)015<3206:OOTSMO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
Save