The 1985 Nevado Del Ruiz Volcano Catastrophe: Anatomy and Retrospection

Home , Armero, Barry Voight, Cotopaxi, Fumarole, Hazard map, Nevado del Ruiz

Journal of Volcanology and Geothermal Research, 42 (1990) 151-188 151 Elsevier Science Publishers B.V., Amsterdam--Printed in the Netherlands

The 1985 Nevado del Ruiz volcano catastrophe: anatomy and retrospection

BARRY VOIGHT

Penn State University, University Park, PA 16802, U.S.A.

(Received December 1, 1989)

Abstract

Voight, B., 1990. The 1985 Nevado del Ruiz volcano catastrophe: anatomy and retrospection. In: S.N. Williams (Editor), Nevado del Ruiz Volcano, Colombia, II. J. Volcanol. Geotherm. Res., 42: 151-188.

This paper seeks to analyze in an objective way the circumstances and events that contributed to the 1985 Nevado del Ruiz catastrophe, in order to provide useful guidelines for future emergencies. The paper is organized into two principal parts. In the first part, an Anatomy of the catastrophe is developed as a step-by-step chronicle of events and actions taken by individuals and organizations during the period November 1984 through November 1985. This chronicle provides the essential background for the crucial events of November 13. This year-long period is broken down further to emphasize essential chapters: the gradual awareness of the awakening of the volcano; a long period of institutional skepticism reflecting an absence of credibility; the closure of the credibility gap with the September 11 phreatic eruption, followed by an intensive effort to gird for the worst; and a detailed account of the day of reckoning. The second part of the paper, Retrospection, examines the numerous complicated factors that influenced the catastrophic outcome, and attempts to cull a few "lessons from Armero" in order to avoid similar occurrences in the future. In a nutshell, the government on the whole acted responsibly but was not willing to bear the economic or political costs of early evacuation or a false alarm. Science accurately foresaw the hazards but was insufficiently precise to render reliable warning of the crucial event at the last possible minute. Catastrophe was therefore a calculated risk, and this combination - the limitations of predication/detection, the refusal to bear a false alarm and the lack of will to act on the uncertain information available - provided its immediate and most obvious causes. But because the crucial event occurred just two days before the Armero emergency-management plan was to be critically examined and improved, the numerous circumstances which delayed progress of emergency management over the previous year also may be said to have contributed to the outcome. Thus the catastrophe was not caused by technological ineffectiveness or defectiveness, nor by an overwhelming eruption, or by an improbable run of bad luck, but rather by cumulative human error - by misjudgment, indecision and bureaucratic shortsightedness. Armero could have produces no victims, and therein dwells its immense tragedy.

"We are all of us fellow passengers on the same planet and we are all of us equally responsible for the happiness and well-being of the world in which we happen to live..." (Van Loon's Geography).

Introduction in 1846 in the Comptes Rendus of the Academy of Sciences, Paris, it captures the Colombian This report begins with Joaquin Acosta's ac- naturalist's observations of a flow of mud that count of tragedy on the Lagunillas; presented swept over the site of Armero in 1845:

"A great subterranean noise was heard in the vicinity regions..." of the Magdalena, from Ambalema to the village of Such ancient traditions encompass the last Mendez... This sudden noise was followed, within a major eruption at Ruiz, known through the considerably smaller area, by trembling of the ground. words of Pedro Sim6n, a Spanish priest who Then, descending along the Lagunillas from its sources in the Nevado del Ruiz, came an immense flood of thick came to the New Kingdom of Granada in 1604 mud which rapidly filled the bed of the river, covered or (Sim6n, 1625; cf. Espinosa, 1986; Vel~squez et swept away the trees and houses, burying men and al., 1986): animals• The entire population perished in the upper part and narrower parts of the Lagunillas valley. In the "It happened then, that on that day, month and year [12 lower part, several people were saved by fleeing March 1595]... there came from this volcano such a sideways to the heights; less happily, others were loud, hoarse, and extraordinary thunderclap, and after stranded on the summits of small hills from which it it three others not so strong, which were heard within was impossible to save them before they died... On ar- a radius of more than forty leagues.., the Spaniards riving at the plain with great impetus, the current of saw that the volcano hurled out a large amount of mud divided into two branches. The much larger one pumice, as big as ostrich eggs and from these down to followed the course of the Lagunillas toward the the size of dove's eggs, sparkling red like iron from the Magdalena; the other, after topping the high divide, forge, which resembled erratic stars. Some fell on them traversed the Santo Domingo valley.•, and hurled and on their horses, which disquieted them no little. itself into the Rio Sabandija, which was thus plugged by And on the side of this mountain which faced the east an immense dam. The danger seemed imminent for a

•.. the waters of the Rio Guali, which wets the founda- flood of downstream lands. Happily, plentiful rain dur- tions of Mariquita, it and its companion which flows in ing the night produced enough water to begin clearing the south, called the Rio Lagunillas, both originating in a passage across this mass of broken trees, sands, the snow which melts from this mountain, ran so full of stones, and stinking mud, mixed with enormous blocks ash that it looked more like a thick soup of cinders than of ice which descended from the mountains in such like water. Both overflowed their channels leaving the abundance that after several days they had not melted land over which they flowed so devastated that for many entirely•.. The terrain covered by debris and mud is years afterward it produced nothing but small more than four leagues square; it presents the ap- weeds..." pearance of a desert or playa, on the surface of which loom up like many islands heaps of broken trees that resisted the impetus of the torrent. The depth of the Armero, which did not exist when Sim6n and mud layer varies greatly, and is much greater toward Acosta made their journeys, was built on the the upper part of the deposit where it often reaches 5 to 6 m. A realistic calculation indicates more than 300 site of these ancient mudflows. Therefore no million tons of muddy material from the flanks of precedent was established when on the evening Volc~in Ruiz..." of November 13, 1985, Nevado del Ruiz ignited to generate the worst volcanic mudflow Of this event: Acosta remarked in an 1850 let- disaster in historic time and the second worst ter to the French geologist Elie de Beaumont volcanic disaster of this century. Its death toll that all of the property which families had ac- ranks fourth in history, behind only Tambora cumulated in years of effort, the fruit of labor of in 1815 (92,000) and Krakatoa in 1883 (36,000), several generations, was lost in a mudflow both in Indonesia, and Mount Pel~e, Martini- which destroyed houses, harvests, flocks, and que, in 1902 (28,000) (UNDRO, 1986). rendered the land barren for a great many Shortly after 21:00 local time, a relatively years: "It is astonishing that none of the in- small magmatic eruption at the summit crater habitants of these villages, built on the of Volc~n Ruiz produced a series of pyroclastic solidified mud of old mass movements, has even flows and surges that turbulently scoured and suspected the origin of this vast terrain, which melted part of the summit's snow and ice cap, occupies an area at least equal to that of the and sent torrents of meltwater, slush, ice and province of the Rh6ne, although ancient tradi- pyroclastic debris in a plexus of sheet and chan- tions testify to the frequent mudflows in these neled cascades over the volcano flanks (Katsui THE NEVADO DEL RU1Z VOLCANO CATASTROPHE: ANATOMY AND RETROSPECTION 153 et al., 1986). In lower channels the flows percent of the region's livestock, thirty percent coalesced and entrained debris, vegetation and of its grain sorghum and rice crops, and half a ponded water to form lahars, the Javanese million bags of Coffee were lost. Lahars buried word used internationally to mean a mud of 3,400 hectares of the agricultured land, damag- debris flow of volcanic origin. ed or eradicated 50 schools and 2 hospitals, On the volcano's fertile western slope, over- destroyed 5,092 homes, 58 industrial plants bank flooding by lahars caused 1,100 fatalities and 343 commercial establishments, and at Chinchinh, Caldas Province, and destroyed damaged the National Coffee Research Center more than 200 houses and three bridges (E. in Chinchin£ About 7,700 were rendered Parra, P., written commun., 1986). Lahars homeless. Total damage exceeded a billion feeding the Rio Guali flowed northward and dollars (U.S. AID, written commun., 1986). then east, destroying homes near Mariquita. Thus "the majestic and silent Nevado del Ruiz, Shortly before midnight, in Tolima Province, transformed now to an active volcano, could not successive lahar waves obliterated Armero resist singing a hymn of destruction; and where (population 29,000), the prime regional there was life and hope of production, there is agricultural center. About 5 vertical kilometers today an ocean of mud as mute witness to the below the summit of Ruiz, Armero became a released fury of nature" (Epoca, Dec. 1985). crypt sealed in lahar mud (Fig. 1). Over 20,000 Yet the eruption was not a surprise, and were entombed and 5,000 more were injured. neither were its effects. Persistent anomalous In the vicinity of the volcano, all roads, fumarolic, phreatic and seismic activity had bridges, telephone lines, power grids and served as precursors for about a year. Colom- aqueducts were damaged or destroyed. Sixty bian workers in hazard assessment and

Fig. 1. The town of Armero following the November 13 eruption of Nevado del Ruiz (photo Steve Raymer, courtesy of National Geographic). 154 B. VOIGHT management were assisted by international ex- plement a geophysical and geochemical pro- pertise. Despite this, the emergency manage- gram for monitoring a probable eruption" ment system failed to avert disaster. (CHEC unpubl, doc., February 4, 1985; Hall, 1990-this volume). Anatomy of the catastrophe In response, a civic committee was formed in Manizales, a city of 350,000 near Volc~n Ruiz. "The narrator ... would have little claim to com- With the support of CHEC, regional corpora- petence for a task like this, had not chance put him in tions and local government, its task was "to the way of gathering much information, and had he not form and support a scientific commission to been, by the force of things, closely involved... This is monitor the volcanic and seismic hazard" (Hall, his justification for playing the part of a historian •. • The present narrator has three kinds of data: first, 1990-this volume). Colombia's Geology and what he saw himself; secondly, the accounts of other eye Mines Bureay (INGEOMINAS) was then con- witnesses; and lastly, documents that subsequently tacted, and on several occasions in the period came into his hands• He proposes to draw on these February 18-27 geological teams from CHEC records whenever this seems desirable, and to employ and the University of Caldas, or from IN- them as he thinks best." (Albert Camus, The Plague) GEOMINAS, visited Ruiz to investigate its activi- The awakening ty. The first articles on the awakening Ruiz appeared in the Manizales newspaper La Patria

"You must picture the consternation of our little on February 21 and 22, and a feature section on town, hitherto so tranquil, and now, out of the blue, March 3 included the full report of the scientific shaken to its core, like a quite healthy man who all of commission of the civic committee and photo- a sudden feels his temperature shoot up and the blood graphs of the crater. At Chinchin~, the only seething like wildfire in his veins .• ." (Camus) seismograph that had been operating in the region broke down on February 23. The Ruiz chronicle begins in November 1984, John Tomblin, a seismologist from the United when after a century of peaceful dormancy, Nations Office of the Disaster Relief Organisa- earthquakes were felt at the Refugio lodge near tion (UNDRO) at Geneva, was in Colombia in the summit of the mountain and climbers March on another mission. Accompanied by reported unusually strong activity in fumaroles two Swiss seismologists, he investigated Vol- within the summit crater. Three significant c~n Ruiz on March 9 at the request of Colom- earthquakes were felt within 20-30 km of the bia's civil defense agency (Defensa Civil) and volcano on December 22, one of magnitude INGEOMINAS and witnessed a 100-150-m vapor 3-4, and over a two-day period, episodes of column above the summit crater. He concluded tremor lasting 5-30 minutes were reported at that "the abnormal activity at Volc~n Ruiz cor- Chinchin~i (Tomblin, 1985a). At Refugio on responds to typical precursory events for an December 22, 65 shocks were felt about every eruption of magnitude". He recommended the 15 minutes between 17:00 and 22:00 (Hall, immediate installation of a portable seismo- 1990-this volume). Snow on Ruiz was covered graph on Ruiz and noted that INGEOMINAS had with a fine ash and sulfur veneer. the obligation to conduct monitoring, using in- Felt earthquakes and strong fumarolic activi- ternational expertise where necessary, and to ty continued into 1985, and on January 6 prepare a hazard map in anticipation of dif- geologists from the Central Hidro-Electrica de ferent types of eruption. The Colombian civil Caldas (CHEC) visited the summit crater Arenas defense was alerted to its obligation to develop and observed that a new, smaller crater had a plan for alarm and evacuation of the populace formed within it. Further investigation led in high-risk sectors, as outlined in an up-to-date them to conclude that authorities should "im- UNDRO-UNESCO Handbook (1986), on Volcanic THE NEVADO DEL RUIZ VOLC,~NO CATASTROPHE: ANATOMYAND RETROSPECTION 155

Emergency Management. An absence of credibility On the surface, INGEOMINAS appeared to be in a reasonable position to prepare a hazard map. "Please answer me quite frankly. Are you absolutely convinced it's plague?" Only three years before, they had published a "You're stating the problem wrongly. It's not a ques- 48-page report on the glacial and volcanic tion of the term I use; it's a question of time." (Camus) geology of the Ruiz-Tolima volcanic complex (Herd, 1982). Based on a 1974 Ph.D. disserta- On March 29 in Geneva, Tomblin expressed tion at the University of Washington, this his concern in a telephone communique to Dr. study identified post-glacial lahars and Michio Hashizume, UNESCO Natural Hazards pyroclastic flows high on the flanks of Ruiz, and Unit, Paris (Epoca, Dec. 1985): "nothing is be- established a dated chronology through tephra ing done about this new activity. It appears correlations and radiocarbon dating. The that the University in Bogota does not have lahars of the volcanic complex were noted to be volcanic expertise, and neither does the in- "more commonly associated with snow-capped stitute for geology and mines (INGEOMINAS). volcanoes", with heat generated by explosive That institute probably carries the responsibili- eruptions cited as the cause of rapid melting of ty in a volcano crisis situation." Hashizume ice. INGEOMINAS had also provided support for subsequently requested Gudmundur Sigval- more recent investigations along similar lines dason in Iceland, representing the World by scientists from Grenoble, France (Thouret et Organization of Volcano Observatories (WOVO), al., 1985). Although volcanic hazards were not to nominate a team of experts in case assistance specifically mentioned, these reports contained was requested by the Colombian government. detailed soil and tephra columns and geological On April 9, wovo telegrammed candidates maps that included historic lahars of 1595 and from Costa Rica, Ecuador and Mexico. The 1845, thus providing the basic building blocks Costa Ricans responded the next day with an of- for hazard evaluaton. However, these publica- fer of three portable seismographs, an elec- tions were founded on foreign expertise, and IN- tronic tiltmeter and dry-tilt equipment. In a GEOMINAS lacked in-house experience in vol- follow-up letter to UNESCO on April 15 canology. Sigvaldason expressed his preference that a On March 20, a local conference on volcanic four-man team be involved, including experts risk was held at the National University in from each country. On April 18, Hashizume sent Manizales. The participants - many of whom invitations to these individuals with the pro- would later form the nucleus of the Volcanic viso that the Colombian government formally Risk Committee - concluded that the possibili- request their services. Meanwhile, at Volc~n ty of an eruption could not be dismissed, risk Ruiz, there were 17 felt seismic events in studies were the responsibility of the state, March, 18 more in April. Sounds interpreted as communities had the right to be informed and rock and ice falls were reported from high in protected, and an emergency evacuation plan the Rio Azufrado. was needed (Calvache et al., 1985). IN- On May 4-7, Minard Hall, Escuela Polit6c- GEOMINAS Report 1937 on the late February nica Nacional, Ecuador, visited Manizales on field examination of Ruiz concluded otherwise: behalf of wovo and UNDRO to evaluate the ac- "the volcanic activity can be considered 'nor- tivity of the volcano and determine what mal' for active volcanoes and does not represent measures had been initiated to monitor it and imminent danger." Nevertheless, they ad- evaluate risk. He reported that "the crater's ac- vocated establishment of a monitoring pro- tivity remains stable but in an abnormal gram. state," with sulfur salts condensing on summit 156 B. VOIGHT snowfields. He was concerned that "no monitor- Robert Tilling of USGS that the INGEOMINAS re- ing activities are being carried out. Portable quest was under consideration (unpublished seismographs, if they exist, are still in Bogota," correspondence), on June 13, Hashizume looked and he suggested that local governments might forward to enrichment of a UNESCO effort by a buy then from foreign sources. "At least four cooperating USGS team. But USGS officials would be needed" (Hall, 1985, Spanish version). ultimately declined to send a technical expert, Hall saw "no attempt to initiate a volcanic-risk claiming prior commitments and studies evaluation despite the fact that good quality associated with recent eruptions in Hawaii and maps exist" and stressed the importance of Mount St. Helens. On July 2, the USGS Deputy making a very preliminary risk map so that Chief for Latin America observed in a note to Defensa Civil would be better prepared, UNDRO, "The opportunity is clear, and it is un- especially for damage reaching the foot of the fortunate that we can spare no one from the Cordillera "where larger towns might be Hawaii or Cascades Observatories... If the threatened by flowage deposits." He referred to volcano is to blow, let us hope that both we and the Cotopaxi, Ecuador, volcanic hazard map as the Colombians were prepared." an example of what should be done (Miller et By early July, responding to the perceived al., 1978), and concluded that the geothermal vacuum in hazard evaluation, a Volcanic Risk department of CHEC was "best prepared and Committee (Comit~ de Estudios Vulcan61ogicos more interested" in carrying out hazard mapp- Communidad Caldense) had crystallized from ing and evaluation. INGEOMINAS ad- the earlier civic committee in Caldas Province ministrators in Medellin were aware of the with the support of the regional government, hazards but "other priorities and the lack of CHEC, the National University in Manizales, funds precluded future work there" (Hall, 1990- the University of Caldas, the coffee and finan- this volume). cial corporations, and concerned citizens The civil defense of Caldas Province had (Medina J., 1986, and oral communications). meanwhile prepared a disaster plan which ap- This organization was confirmed by govern- peared to Hall to be "excellent and completely ment decree in August. The Comit~ initiated adequate. However, once the preliminary programs for investigation, local risk planning hazard map is prepared, it will be necessary to and public education. On July 7 they acknow- review the plan and to adjust it in accordance ledged INGEOMINAS for obtaining seismographs with the map. It is recommended that the plan from ISA, and expressed the futile hope "that be approved" (Hall, 1985, Spanish version). He in the next few days INGEOMINAS will turn over made no mention of a similar plan for the these instruments to our project" (Hall, 1990- neighboring Tolima Province in which Armero this volume). The Mayor of Manizales and lies, although he spoke briefly by phone to the Governor of Caldas submitted an official re- national director of Defensa Civil. "All groups quest for technical assistance and equipment to were interested" in the possibility of col- the Swiss Disaster Relief Corps, and an UNDRO laborating with a volcanology team sent from telex on July 10 to the Swiss Seismological Ser- abroad. vice supported the Colombian need for On May 30, INGEOMINAS made a request to assistance in seismic operations and interpreta- the U.S. Geological Survey for technical exper- tion. tise and for geophones and cable needed to On behalf of the Swiss Disaster Relief Corps operate in-country seismographs on loan from and Swiss Seismological Service, Bruno Mar- Interconexi6n El~ctrica S.A. (ISA). The parts tinelli arrived in Manizales on August 8 to were provided at UNDRO expense and shipped assist Colombian scientists with seismic by diplomatic pouch on June 26. Informed by monitoring. His equipment included a portable THE NEVADO DEL RU1Z VOLCANOCATASTROPHE: ANATOMYAND RETROSPECTION 157 three-component short-period event recorder, (1) a team of volcanologists from different and MEQ-800 drum recorder used as a one- countries; component continuous monitor to supplement (2) equipment and instruments for measure- the event recorder and the INGEOMINAS net- ment; work, a one-component system with tape (3) training for Colombians in their country or recorder mostly used to collect data on a at volcanologic centers of other countries; glacier, and Mark geophones (B. Martinelli, (4) exchange of information and experience written commun., 1986). Martinelli was ably concerning similar phenomena. Everything supported and assisted by CHEC personnel and could be provided in a short time ... To ac- by authorities of the Caldas region, and recalls complish this project, it is necessary that the the following (written communication): Colombian government file a formal petition to "On my arrival, four portable seismographs UNESCO as soon as possible." Despite this operated by personnel coordinated by IN- urgency, the letter was apparently sidetracked. GEOMINAS were in operation since July 20th in Nearly two months later it resurfaced, attached the region around the Nevado. However, they to an August 21 note from an employee of the did provide only little information of relevance Ministry of Education to the Governor of due to their non-optional locations." (They were Caldas Province: "In this communication... relocated in September.) "It was only at the end the delegate to UNESCO requested that you of August, and due mostly to the contribution of should petition UNESCO in order to stop the Juan Duarte (Institute de Los Andes) that the reactivation of the volcano (sic) (El Tiempo Oct. seismic system could be considered to be 25, 1985; Epoca, Dec. 1985)." Within a few suitable for monitoring the volcano at all... days, the letter was in the hands of the Volcanic Cooperation between myself and the persons Risk ComitY; but by then the volcano was forg- responsible for the seismic network was always ing a schedule of its own design. characterized by cordiality and reciprocal consideration. In particular the cooperation with Juan Duarte was very intensive and pro- The credibility gap closes fitable." The Comit~ and INGEOMINAS were now "(This) marked one might say, the end of the first period, that of bewildering portents, and the beginning operating seismographs independently but the of another, relatively more trying, in which the perplex- data were not amalgamated for rapid process- ity of the early days gradually gave way to panic.., our ing and complete interpretation. The IN- townsfolk realized that they had never dreamed it possi- GEOMINAS data were processed in Bogota. ble that our little town should be chosen out for the Although background levels were unknown, scene of such grotesque happenings ... in this respect seismic records indicated about 5-20 countable they were wrong, and these views obviously called for revision . .. And it was then that fear, and with fear earthquakes daily, with as many as 40-50 dur- serious reflection, began." (Camus) ing swarms and occasional tremor (Cu~llar and Solano, 1985; Fig. 2). Maximum magnitude was E1 Ruiz commanded attention at 13.30 on about 3.5 (D. Harlow, oral commun., 1986). September 11, when a strong phreatic eruption On June 26, the same day that the geophones occurred at the summit crater, lasting for were shipped by USGS, the Colombian delegate perhaps 7 hours. It had been preceded by 15- to UNESCO wrote to the Minister of Colombian minute episodes of tremor every 80 minutes or Foreign Affairs, called attention to the Tomblin so since September 6 (Fig. 2; Harlow et al., and Hall UNDRO reports and noted the follow- 1986). Violent steam explosions ripped pre- ing: "UNESCO has contacted several interna- existing ash and blocks from the volcano tional organizations and is able to offer Colom- throat, deepening and steepening the walls of bia: the summit crater (SEAN, 1985; Tomblin, 158 B. VOIGHT

ERUPTION I I

TREMOR III H II liili Ilill II I Hi

50 LU AI " Refugio~ z f" >- ,_1 s

o 0 I - I

28-

24-

>- ¢9 a:¢9 ~ 20

Z UJ ~ 16 J tkl ~O 1 J I'-- ¢~X 12

~ 8

(..) 4 j

t 1 JULY AUG SEPT OCT NOV 1985 Fig. 2. Volcanic events, seismic events and energy release at Nevado del Ruiz (data after INGEOMINAS, D. Harlow and J. Zollweg).

1985a). Lithic blocks were cast kilometers from ticularly intense and the monitoring seismic the crater and ash began to rain on Manizales system was operational by this time. However, (Fig. 3) and Chinchin~i, 25 km distant. In mid- this strong activity only started around the afternoon, ice, firn and rock avalanches were beginning of September. On the 10th of the dislodged from the headwall of the Rio same month, at a meeting organized by the Azufrado, probably reflecting the combined ef- emergency committee at which not only myself fect of seismicity and fluid pressures associated but also INGEOMINAS, the Governor of Caldas with snowmelt and water regurgitated from the and the Lord Mayor (Alcalde) of Manizales were summit crater. At 18.30 a lahar extended more present, the seriousness of the situation was than 20 km down the river, eroding through the discussed. It was on this occasion that concrete road link between Manizales and Murillo (E. proposals were formulated for a complete and Parra P., written commun., 1986). Inhabitants long-term systematic monitoring of the vol- of the valley were alerted but no evacuations cano, and it seemed to me that all of those pre- occurred. Martinelli recalls the situation (writ- sent were convinced of the necessity of this ten commun., 1986): course of action..." "The seismic activity prior the the phreatic Martinelli expressed concern that the IN- eruption of the llth of September was par- GEOMINAS seismograms were sent to Bogot~ THE NEVADO DEL RUIZ VOLCANO CATASTROPHE: ANATOMY AND RETROSPECTION 159

Fig. 3. Nevado del Ruiz in September 1985, seen from Manizales in Caldas Province. At this time the population was in a state of "fear, if not panic". The focus on Manizales may have retarded local emergency management in neighboring Tolima Province, where even greater risk existed (photo Steve Raymer, courtesy of National Geographic). without being processed; further "The danger dangerous" in an attempt to calm the popula- of the lahar was not only evident, even to the tion. layman, but after the eruption of the llth of On September 15 the Governor of Caldas con- September there were no longer any doubts ducted a personal inspection of Ruiz (El Tiem- that some measures must be taken in order to po, Sept. 18, 1985) and organized a meeting on deal with this danger." September 16 among the governors of Caldas, Thus on September 13 the Volcanic Risk Tolima, Risaralda and Quindio and the Direc- Comit~ in Manizales issued a public warning of tor of INGEOMINAS (El Tiempo and El Espec- serious risk of additional avalanches of rock tador, (Sept. 15, 1985). The governors "agreed and ice and joined with Defensa Civil and the that each province would autonomously Governor of Caldas in recommending evacua- manage their emergency plans" (La Patria, tion along rivers fed by the snow and ice of Ruiz Sept. 17, 1985) and appealed to national (El Espectador, Sept. 13 and 14, 1985). These government and international experts for statements received considerable media atten- assistance (La Patria, Sept. 18, 1985), whereas tion although La Patria (Sept. 13) reported on INGEOMINAS emphasized the high priority of the front page that "Ruiz activity is not monitoring (Parra and Cepeda, 1990-this 160 B. VO1GHT volume). An aviation advisory suggested that if a lahar was generated on the Lagunillas..." pilots should not fly within 50 miles of Ruiz (El (El Tiempo, Sept. 18, 1985). INGEOMINAS con- Tiempo, Sept. 17, 1985) On September 18 the cluded that the dam, which was not related to Bogot~ newspaper El Tiempo published a color volcanic activity, did not appear to be in danger photograph of Ruiz on the cover page, and gave of failure "under normal river conditions"; at extensive coverage to the developing crisis. the same time, a drainage plan was recom- On September 17 the mayor of Armero mended but construction was never started (E. reported to authorities that his town was Parra P., written commun., 1986). The actual threatened by a landslide-dammed lake at reservoir volume at Cirpe was perhaps 250,000 Cirpe, 12 km upstream on the Rio Lagunillas m 3 (R. Janda, oral commun., 1986). and containing estimated half a million cubic Meanwhile on September 16, Defensa Civil, meters of water, and that flood hazards existed INGEOMINAS, Ministry of Mines, CHEC, Caldas also on other rivers such as Rio Guali and Rio government officials and coffee-industry repre- Recio, due to geothermally enhanced ice sentatives met in Bogot~ to consider emergency melting on Ruiz. Regarding the Lagunillas plans for Ruiz: "a 25 percent chance of an erup- dam, "an increase in the volume will bring the tion was estimated," and Defensa Civil total destruction of Armero. The town has "declared a state of alert" (El Tiempo, Sept. 18, already had two previous floods with serious 1985). consequences for the population" (La Patria, The September eruption had clearly slammed Sept. 21, 1985). (Floods occurred in 1935 and shut the credibility gap. On September 17, a 1950 from torrential rains.) meeting of emergency and civil-defense direc- The mayor identified specific barrios border- tors convened at the request of the government ing the river within Armero to be at risk to to form a national-level emergency committee "tragedy" and named an 8-man emergency to coordinate a plan a response to any further committee of civic leaders to request immediate activity on Ruiz (El Tiempo, Sept. 18, 1985; action from the President of the Republic. The Hall, 1990-this volume; Herd and ComitY, committee, which included the local chiefs of 1986). Defensa Civil developed a management Defensa Civil and Red Cross, stated the follow- plan and attempted to identify what remained ing (La Patria, Sept. 21, 1985; cf. El Tiempo, to be done, including procedures to carry out Sept. 18, 1985): "We are concerned about the evacuation. Red Cross assumed responsibilities situation created by Volc~n Nevado del Ruiz. for emergency communications and disaster This is unpredictable, but we can see that response. INGEOMINAS was assigned national critical danger threatens the population. For responsibility for monitoring and hazard this reason we request that the Rio Lagunillas assessment, generally following UNDRO's time- dam be dealt with, because it represents a time worn suggestions of March 9. These agencies bomb. The problem of Nevado del Ruiz is well met with the Colombian Congress to obtain known to the government and to public opinion, funding for the work required. An inter- as scientists have issued warnings about it. For parliamentary committee involving the four this reason we urge the government to take the provinces influenced by Ruiz was formed to appropriate measures to prevent inundation by oversee state aid for scientific work (El Espec- the river, as occurred on previous occasions tador, Sept. 19, 1985). with consequences disastrous to the population. A preliminary draft of a volcanic hazard map Today the possibility of a catastrophe is greater was scheduled for October 7 and visual observa- because of 'superpopulation' of the city." The tion and monitoring of Ruiz was to be expand- director of Red Cross for Tolima Province sup- ed. After September 20, six INGEOMINAS ported this view: "Armero... might disappear geologists and four University of Caldas in- THE NEVADO DEL RUIZ VOLCANO CATASTROPHE: ANATOMY AND RETROSPECTION 161 structors and some students joined forces at a tant to a geothermic company.., a member of retreat at Termales del Ruiz, on the flank of the the monitoring committee made phone contact volcano not far from Manizales, to collaborate with Franco Barberi in Pisa on September on the construction of the hazard map (Parra 13th. I also took the opportunity to inform him and Cepeda, 1990-this volume). On September of the seismic recordings made, and it wasn't 23, before returning to Switzerland, Martinelli difficult to convince him of the necessity of his (written commun., 1986) had a chance in presence in Manizales" (B. Martinelli, written Bogot~ to view a first draft of this hazard map: commun., 1986), cf. CHEC, 1983). The Italian "Dario Moschera illustrated to me the criteria team arrived in October. • .. the danger of lahars, especially along the Meanwhile, the credibility gap had closed suf- Rio Azufrado and Lagunillas, was the guiding ficiently for the USGS to release Darrell Herd, theme of this work." The Instituto Geogr~fico the administrator in Reston who had studied Agustin Codazzi was asked to establish a Ruiz for his Ph.D. dissertation. Supported by geodetic polygon and provide aerial the U.S. AID Office for Foreign Disaster photographs and maps. A chief of seismology Assistance (OFDA), Herd spent the week of was drawn from the National University of September 20-27 assisting Colombian scien- Bogot~ by INGEOMINAS, and regional emergen- tists in Manizales to evaluate past eruptions as cy committees were established to coordinate a basis for future prediction and to plan in- local response planning. The existing Volcanic vestigation programs. On September 23 he Risk Comit~ (involving CHEC and regional gave a public lecture to help dispel what was universities) in Manizales was assigned respon- then perceived to be "the unnecessarily great sibility for scientific studies at Ruiz, with input anxiety of the population of Manizales," since from INGEOMINAS. Nevertheless on September UNDRO had been informed that the population 24 the Congressional representative for Caldas was in a state of "fear, if not panic" as a result "lamented that requests for help sent to IN- of the September 11 eruption (Tomblin, 1985a). GEOMINAS, to Ministry of Foreign Relations El Espectador, Sept. 28, 1985 reported that "an (and others).., had not been successful... The eruption would only affect an area within ap- national government was acting with great proximately 10 km of the summit." uncertainty" (La Patria, Nov. 17, 1985; Hall, Manizales was indeed the hub of activity near 1990-this volume). In response, the Director of Ruiz, being the one large city close to the Defensa Civil emphasized the regional meet- volcano. However, Manizales was removed and ings that had been held in July, and the in a different province from most of the main Minister of Mines and Energy said that scien- rivers at risk, which flow eastward toward the tific and mitigation studies would be completed Magdalena. In retrospect, this focus on Mani- "before the end of October" (La Patria, Sept. 26, zales may have retarded local emergency 1985; El Espectador, Sept. 27, 1985; Hall, 1990- management in neighboring Tolima Province, this volume). where even greater risk existed. This view is Formal petitions for assistance had been sent sustained by Martinelli (written communica- to UNDRO, UNESCO, USGS and others. In Mani- tion): "During the whole period of my stay zales, "the eruption of the llth September con- there were no mutual contacts or cooperation vinced the monitoring committee of the necessi- between the region Caldas, where I was, and ty of immediately having a volcanologist on the Tolima. Only on September 13th, i.e. after the spot who possibly had prior knowledge of the eruption of September llth, was a meeting held Ruiz. The choice did not prove difficult seeing at the Defensa Civil in Bogota, organized by that Franco Barberi had occasion to work for General De La Cruz, with representatives of several months on the Ruis in 1983 as a consul- the authorities of Caldas and Tolima, of the Red 162 B. VOIGHT

Cross, and the Defensa Civil. As far as I could INGEOMINAS. I was never able to figure out the see, even after this meeting cooperation be- reason for this conflict nor to clarify what tween the two regions was insufficient." responsibilities each institution actually had. A geochemist from New Zealand aquired In any case, this rivalry had a strong influence fumarole-gas samples between September 22 on the whole phase of emergency preparation." and 30 (partial analysis revealed significant Tomblin therefore urged closer cooperation bet- SO 2) and John Tomblin of UNDRO took over the ween the two groups in Manizales to permit all advisory role from Herd and Martinelli from local earthquake data after September 26 to be September 25 to October 2. He met with na- interpreted rapidly and as fully as possible, and tional leaders of civil defense and INGEOMINAS, through UNESCO, coordinated the return of participated each evening in scientific and ad- Minard Hall from Ecuador to guide the ministrative meetings, and reviewed the work cooperative effort. He also arranged a mission of some local management and vulnerability- by Rudolfo Van der Laat from Costa Rica to assessment groups. Several hours each after- carry out ground-deformation measurements. noon were spent working on the hazard maps. In a mission report dated October 9, Tomblin Tomblin (1985a, unpublished correspondence) (1985a) noted that eruptions producing ash- learned that the seismology network needed falls and mudflows were more likely than other prompt attention. He attempted to acquire types of eruption and observed that "scenarios radio-linked seismographs to use in short-term which should be envisioned for the purpose of prediction, since the seismograph net in opera- hazard zoning and emergency planning (in- tion relied for recording on rotating drums of clude) rapid melting of the glacier;" further, smoked paper which had to be retrieved every "the area devastated in 1845 has a present-day day in the field. Requests were made by phone population of the order of 20,000." He also com- and telex to the Volcano Observatory in Mar- mented that the reactions of scientists and tinique, the University of Paris, the USGS, the public safety authorities were greatly stimu- Swiss Federal Seismology Institute, the Swiss lated by the September 11 eruption, so that Disaster Relief Corps and the Swiss Am- "'within the very near future, the necessary bassador in Bogota. The problem had not been measures will have been taken to protect the resolved by the time of his departure. population". Martinelli expressed a similar opi- Equally important, the Risk Comit~ and IN- nion; though well aware of several obstructions GEOMINAS were still not fully sharing seismic to preparedness, on leaving Manizales on data. The Comit~ (through Martinelli and September 24 he "was convinced that CHEC) had several stations, but an insufficient everything would be done to limit the possible number for hypocenter location study. IN- damage to a minimum." GEOMINAS had four stations (they added another on October 23), but data had to be mail- Girding for the possibility of battle ed to the National University of Bogot~ for interpretation; not only was short-term predic- "People in town are getting nervous, that's a fact tion infeasible, but no seismic data or inter- • .. And of course all sorts of wild rumors are going around• The Prefect said to me, 'Take prompt action if pretations were actually issued by INGEOMINAS you like, but don't attract attention.' He personally is until October 7, despite operation of the system convinced that it's a false alarm." (Camus) since July 20. Martinelli (written commun., 1986) reported "cordiality" between seismic There was general satisfaction that the need specialists from the two groups, but "during the for uniting all available national resources to whole time of my visit I noted a pronounced meet the volcanic threat was finally recognized. rivalry between the emergency committee and Political fences had been mended and pre- THE NEVADO DEL RUIZ VOLCANO CATASTROPHE: ANATOMY AND RETROSPECTION 163

disaster planning was at last progressing at a "volcanic terrorism" (La Patria, Oct. 4 and 5, reasonable pace and in the right direction. 1985). On October 7 La Patria noted in a cover Minard Hall, who returned to Manizales Oc- story that publication of a volcanic-risk map tober 2-14, set up a hypocenter-location pro- would cause real estate "devaluation" (cf. La gram and assisted in the coordination of Patria, Sept. 18, 1985). Nevertheless, the seismic data and the completion of the volcanic preliminary version of the hazard map (scale hazard map. 1:50,000) was presented in a press conference There were still some countercurrents. The on October 7 attended by the Minister of Mines, Director of the Geophysical Institute of the the Governor of Caldas (not Tolima?) and direc- Andes reported in Magazin 8 Dias (Sept. 1985) tors of INGEOMINAS, Defensa Civil and other that "Today ... nothing is happening on the dignitaries (Fig.4; cf. Parra and Cepeda, 1990- volcano that threatens the inhabitants... If it this volume, fig. 2). "The map showed that ex- is necessary to declare an emergency, the In- tensive areas.., are threatened (and) will help stitute will do it. To do so before would only various civil defense organizations design plans alarm the populace without reason" (Hall, for evacuation.., some cities such as Armero 1990-this volume). Likewise in Manizales, in would have 2 hours to evacuate..." (La Patria, response to conflicting opinions, the Chamber Oct. 8, 1985). But only ten copies of the map and of Commerce feared that irresponsible repor- accompanying explanation were prepared and ting "will cause economic losses" and the Arch- distributed (Hall, 1990-this volume), although bishop criticized the media for the spread of INGEOMINAS staff also met with officials of

\ Fresno• \

Paleslina• f Chinching. Sant 4Libano u Lava FlowHazard MudflowHazard m • Lerida PyroclasticFlow Hazard High ~ I~ Moderate Ashfall Hazard High I~ ~ Moderate Low Angle8last Hazard High ~ [~ Moderate 20 Km. Limit i 5 10 Mes I I C 5I 11 115 K,Iometers Ibagu6

Volcanic hazard map of Nevado del Ruiz. [INGEOMINAS]

Fig. 4. Volcanic hazard map of Nevado del Ruiz (November 1985 version), showing locations referred to in text. 164 B, VOIGHT

Armero, Mariquita and other towns in Caldas overflights of territory in the hazard zones, and and Tolima and informed them of the potential in late September and October compiled a field hazards (F. Zambrano O., cited in Herd and count of the population at risk along most ma- Comit6, 1986). The accompanying INGEOMINAS jor river systems fed by Ruiz. By October 29 the report (Cepeda et al., 1982) referred to lahars as Azufrado and Guali still remained to be assess- follows: ed (Defensa Civil documents). Volcano- awareness programs were initiated in schools "The presence of old mudflow deposits in some river valleys indicates several possibilities for such events in Caldas (El Espectador, Sept. 18, 1985). A during a large eruption. The probability for this type of meeting with scientists and Tolima govern- event is high for the rivers indicated on the hazard map, ment representatives on September 30 discuss- and they could occur during any eruption - therefore ed the need for "persons with radios in strategic the probability is 100%. The magnitude is closely places along the rivers to help prevent a related to the size of the eruption and the availability of water.., for the areas of high hazard the flow could be catastrophe" (El Espectador, Oct. 1, 1985). 25 m deep, and at curves and narrows it could reach 50 Radio facilities were augmented by fixed and m." portable communication devices, and supplies and equipment such as ambulance tires, The report was accurately referred to in emergency rations, boots and uniforms were media articles. A generalized version of the purchased. Defensa Civil representatives of the map was published in color on the front page of provinces at risk met with national officials to the Bogota paper El Espectador (though with provide regional input for emergency-response some errors in map symbols), and articles men- planning, and then conducted meetings with tioned "a 100% probability of mudflows... local officials in Armero, Chinchin~ and Honda with great danger for Armero, Mariquita, Hon- to disseminate risk and preparedness informa- da, Ambalema and the lower part of the Rio tion. For example, a presentation on "Two Chinchina (El Espectador, Oct. 8, 1985) and Natural Catastrophes" was announced for Oc- "lahars and floods are inevitable... Armero tober 1 and 2 in Ibagu~ "to illustrate to mayors would be evacuated in two hours without and other officials in Tolima province the pro- danger" (El Tiempo, Oct. 8, 1985). per response to be followed in the event of a The map was criticized by some government volcanic catastrophe" (El Espectador, Sept. 30, officials in Bogota for "being too alarming" 1985). (Cepeda et al., 1986). The mayor of Armero Much of the information transfer to the local stated in a magazine interview that many peo- level was coordinated by the emergency com- ple in Armero now did not know "whether or to mittees in each province, which included stay leave" and lamented that the local representatives from the Governor's office, emergency committee "did not have the Defensa Civil, Red Cross, INGEOMINAS, police, necessary information or financial resources to military, health organizations, and others. do anything in the event of a catastrophe. For These committees contacted villages to inform this reason, the people have lost confidence in them of the need for preparedness and encour- the veracity of the information and have com- age the development of local evacuation plans. mended their fate to God" (Cosnigna, Oct. 15, Defensa Civil volunteers talked directly to local 1985; Hall, 1990-this volume). residents in some risk areas, including village Meanwhile, Defensa Civil was hard at work and riverbank squatters, telling them about (numerous Defensa Civil documents; P. Bolton civil defense and the current risks from Ruiz, and D. Mileti, written commun., 1986): in and explaining disaster preparedness. One September they assembled risk information by such announcement read as follows: THE NEVADO DEL RUIZ VOLCANOCATASTROPHE: ANATOMY AND RETROSPECTION 165

"The Regional Emergency Committee of C aldas notifies ception of an eruption: "we should consider that you that in view of the danger that exists due to melting the communication system may be damaged in of snow, which causes mudflows or a rise in river levels, the initial moments of eruption." Moreover, the margins of the rivers Molinas, Nereidas, Rioclaro "taking in consideration that lahars are the and Chinchin~i MUST BE EVACUATED and such an EVACUATION must be accomplished to a minimum most dangerous volcanic phenomena and that height of 50 meters above the river IMMEDIATELY." there is a great possibility for them to occur in any type of eruption.., it is advisable to give The notifier's name, the notified, the address high priority to a communication system that and date were also indicated (Defensa Civil will allow an alert to be rapidly sent to people documents). living in areas exposed to the risk of lahars. It Some documents suggest and intentional ef- is necessary to identify a place of refuge in each fort to build public confidence in order to of these towns, and to inform the community so minimize the chances of public panic if a that the people will know where to go in case of disaster occurred. P. Bolton and D. Mileti (writ- an alert." On October 22 the report was ten commun., 1987) subsequently found no delivered to local government and emergency evidence that fear was a byproduct of the effort managers, and left with the Italian Am- to inform those at risk. In September, some bassador for official transmission to the govern- members of the Bogot~ legislature had criticiz- ment in Bogota. By October 31 a complete gas ed both INGEOMINAS and Defensa Civil for caus- analysis on a Ruiz sample was performed in Ita- ing unnecessary fear by their hazard evalua- ly and the results, which verified a magmatic tion and planning efforts (La Patria, Nov. 17, origin for the gas, were telexed to Colombian 1985). But in retrospect, perhaps more fear authorities (Barberi et al., 1986, 1990-this would have been desirable at some locations to volume). encourage an appropriate public reaction. Between October 21 and November 3, with On October 10, in a continuation of Tomblin's support from UNESCO, Van der Laat (1985) con- quest for equipment, Hall sent the following ducted ground deformation monitoring, though telex to Robert Tilling, USGS, "Seismic delay it may be recalled that the Costa Ricans had processing 24 hours makes telemetering been available since April. Although four dry- desirable. On the 16th, Hall advised UNDRO by tilt stations were established about the north- phone that the hazard map had been issued to west side of the volcano and two electronic government authorities, and that although tiltmeters were installed, a retrospective look seismicity remained abnormal, serious moni- at the data "yielded erratic trends not easily toring problems had been encountered (Hall, reconciled with even a small intrusion" (Banks 1990-this volume). et al., 1990-companion volume); further, "the Also on the 16th, an advisory team represen- electronic-tiltmeter data are not useable ting the National Volcanological Group of Italy because the tiltmeters were removed nine days arrived in Manizales to advise on eruption prior to the eruption before settling-in drift had hazards. They reported significant deficiencies stopped." The "reliable" dry-tilt data showed in the monitoring program, particularly with no pre-eruption tilt, perhaps because the source regard to hypocentral location capability, magma was already emplaced or was too deep telemetry, and semiautomatic data processing to detect (Banks et al., 1986, 1990-companion (Barberi et al., 1985): "to this point in the mis- volume). Nevertheless, a consistent tilt pattern sion (20 October) we have only the analysis of implying deflation emerged from background at seismic data on 26 September." The also em- one station (SEAN, 1985), and although in phasized the need for a communication system hindsight considered untrustworthy (Banks et which could provide immediate alert at the in- al., 1990-companion volume), "a few days of tilt 166 B. VOIGHT observations might have provided unchecked ed in an economic crisis, President Betancur misleading information" (Barberi et al., 1986). came under political attack for his November 6 On October 29, the Tolima Emergency Com- decision to send troops against guerrillas who mittee met for the third time at the Red Cross had captured the Palace of Justice in Bogota. office in Ibagu~. The INGEOMINAS represen- The bloody assault left 100 dead, including 11 tative reported that he and other committee supreme court justices (Time, Nov. 25, 1985). members had visited 12 municipalities in Reacting quickly, the State Department and northern Tolima to provide instructions on OFDA concluded not to send U.S. government volcanic risk and advice on precautions for the employees into a potential hornet's nest; White benefit of 4,380 residents. Emergency com- and Harlow were advised to unpack their bags. munication systems were then reviewed by Red On November 6, most INGEOMINAS geologists Cross and the hazard map was presented by IN- working on Ruiz left Manizales in order to com- GEOMINAS, along with hazard probabilities. plete reports at their home offices in other pro- The local emergency plans of Herveo and vinces (Hall, 1990-this volume). Because of the Libano were praised, with the suggestion that Palace of Justice crisis, public presentation of these could be useful models for other the revised hazard map (scale 1:100,000) and municipalities. The meeting log shows that the report was postponed from November 12 to 15 Governor's representative applauded this work (Fig. 4; cf. Parra and Cepeda, 1990-this and suggested that a meeting be held in Libano volume). The November report provided further to establish standards and to improve the local details on the lahar hazard (Cepeda et al., plans in all municipalities at risk. A meeting of 1985b): mayors was therefore scheduled at Libano for "Occurrences of this type of event in these rivers is ex- November 15, to be preceded by an oganiza- tremely high (100%) during eruptive phases of any type, tional meeting on November 13 in Ibagu~. On varying only in magnitude... In the case of an eruptive October 30 a telegram from the Governor's re- event of great proportions (similar to 1845 or larger in presentative and a work order signed by the magnitude), the hazard zone has been evaluated assum- Governor went to the mayors in each Tolima ing a mud thickness of 50 m over normal river level... the maximum assumed thickness has been calculated municipality within potential reach of Ruiz, in- on the basis of the measurement of deposits of mudflows forming them that their emergency plans were observed in the rivers Guali (at Santa Ines) and to be critically reviewed at the Libano meeting. Azufrado-Lagnnillas (near Liban). When the river Meanwhile, in response to Hall's October 10 leaves the canyon, reaching the Rio Magdalena flood telex (but over four months since a plea for plain, spreading occurs, causing a thinning and an increase in the area affected. As an example of this, we technical assistance had been made by IN- have the mudflow of 1845, which near Armero had a GEOMINAS), the USGS had proposed to OFDA the thickness of 8 m, with the area of influence extending to installation of an array of six telemetered the Rio Sabandija and Rio Magdalena... In the Rio seismographs at Ruiz. OFDA declined support, Chinchina are several mudflow deposits, and in lower and weeks were then lost in interagency reaches of the confluence of the Rio Claro and Molinas. negotiations whether the proposed cost should On this basis, the hazard zones were delimited with great detail for the towns of Armero, Honda and Mari- be of the order of $40,000 of $10,000. With the quita." (though not for Chinchin~). lower figure finally agreed upon, USGS seismologists Dave Harlow and Randy White were The detailed map for Armero implied that ready to depart on November 7 with a single most of the population would have to travel a telemetered seismograph (D. Harlow, oral com- distance of over a kilometer in order to reach the mun., 1986). But in a sense symptomatic of the edge of the zone of inundation. The Italian team entire tragedy, the Colombian government pro- had written of the need to identify a place of duced other matters of concern. Already engulf- safe refuge in each town; but the Armero map THE NEVADO DEL RUIZ VOLCANOCATASTROPHE: ANATOMY AND RETROSPECTION 167 implied that no such place existed there and lithic ash was deposited within a distance of any evacuation therefore would be a major about 50 km east of the volcano and muddy rain enterprise. oxidized metal roofs. Punctuated with small ex- On November 10, three days of continuous plosions, tremor then characterized the seismic 2-3-Hz tremor began at E1 Ruiz. At 78 decibels record until 21:08. it was clear but less pronounced than during The regional Red Cross and Defensa Civil of- the September eruption. Nevertheless, during a rices were soon informed that Ruiz was erup- regular meeting of the Comit6 in Manizales on ting. Cr. Rafael Perdomo S., Tolima regional the 12th, several participants voiced concern civil-defense coordinator, reported this informa- about monitoring and public education pro- tion to national headquarters and was in- grams (Hall, 1990-this volume). No analysis of structed to warn local Defensa Civil stations seismicity had been issued since October 10, (Fig. 5). Soon afterward, between 17:00 and and ISA now wanted their seismographs return- 19:30, he joined the Regional Emergency Com- ed. Neither ISA nor INGEOMINAS could maintain mittee for Tolima at their previously scheduled the system, and operations were being sup- meeting at the Red Cross office in the provin- ported by the Comit~ and CHEC. INGEOMINAS cial capital of Ibagu6. The key regional decision was asked to leave them at Ruiz at least until makers were thus already assembled when Red February. On November 11 INGEOMINAS again Cross Defensa Civil officials reported that had claimed that Armero could be evacuated in volcanic ash was falling in northern Tolima, 2 hours without danger - ironically, this state- along with heavy rain. Minutes of the meeting ment was published November 14 in El Tiem- indicate that the police representative was "im- po. mediately asked to communicate with central headquarters in order that telex alerts to prepare for mudflows and floods might be sent Day of reckoning to all police stations near the rivers Guali, Azufrado, Lagunillas and Recio." This was ac- "A picture rose before him of the red glow of the pyres complished via a hand radio in the presence of mirrored in a wine-dark, slumbrous sea, battling tor- the committee and a similar advisory was ches whirling sparks across the darkness, and thick fetid smoke rising toward the watchful sky. Yes, it was radioed by Red Cross to their municipal sta- not beyond the bounds of possibility..." (Camus) tions. In both instances, instructions were made to "sound the alarm - if necessary - in Nevado del Ruiz erupted with little short- the lower reaches of the rivers and at the local term warning on November 13, following Red Cross headquarters." almost a year of precursory activity. The rise in The meeting was then officially called to rate of seismic energy released in October and order, and began with a reading of minutes of early November was only slightly larger than the previous meeting (Defensa Civil docu- background level, in contrast to the sharp ments). Recognizing the possibility of a develop- premonitory increase of energy preceding the ing emergency, discussion first focused on a September 11 eruption. Gas samples were col- suitable location for an operations center. The lected from summit crater fumaroles on the police headquarters suited this purpose because 12th but visual observations yielded no signs of it provided facilities for communication with impending eruption. remote sites. Following a detailed discussion, The events of November 13 began in an agenda was then developed for the forthcom- midafternoon at 15:06 with a phreatic eruption ing Libano meeting on November 15. As this and its associated seismic signal, lasting about meeting was to be inaugurated by the Gover- a quarter of an hour (Harlow et al., 1986). Fine nor, preparations needed to be thorough: it was 168 B. VOIGHT

eruptions

15:06Phreatic 21:08 Plinian Seismograph National Emergency Comlt(; 21:10 - 21:37 with Observer RING RC NP Army Health Others Volcanic Risk Comit=;

nfotify towns] ~ 16:00 eruption J 21:30 Regional Emergency Comite" (Tolima) / I Other Regional Emergency Comit~s I (Caldas, Rlearalda, Ou ndio) RC 1{ N~ Army Health Ibague Meeting 17:00-19:30 I I Ill A17:00T 17:00 ? T Ira I°"~t ] [15:04 eruption] 16:00 ( [t/cod~o,.#,o] I J [21:08 eruption] ( 23:00 • ~u..=~j t'~19"30 Q19:3q r cond/t/on] 1:45 ,B • Mun/k) 21:45 vac?] 6 23:0( 21:30-22:30 22 Caldas Health Church Othem Local Comit~e Local Emergency Comlt~ (Armero) other towns

[22:30 ovac? ] River Observer LJbano,Mud//o, Amba/s,"a R.o A ero 18:00-23:00 I Public [be ca/m] The Public r21:08 eruption] I Radio Priest ~ L be calm J 18:00 [be ca/,.]

Communication 16:00 - Approximate time of communication • OK O ~O NotUncertainoK [ ] Message Q Element in communication DC - Defensa Civil ING - Ingeominas RC - Red Cross NP - National Police Gov - Governor's Representative Pres - President's Representative RA - League of Radio Amateurs

SIMPLIFIED MANAGEMENT SYSTEM AND COMMUNICATION LINKAGES 13 NOVEMBER 1985

Fig. 5. Flow diagram of management system and communication linkages for November 13, 1985, emphasizing connections between national, regional and local authorities, volcano observers, and the public. Heavy link lines connect organized groups; thin links connect individual elements such as Defensa Civil offices. Dot symbols indicate effective (OK) or ineffective communications (not OK), with approximate times and abbreviated messages. Details are given in text. THE NEVADO DEL RUIZ VOLCANO CATASTROPHE: ANATOMY AND RETROSPECTION 169 to include summaries by INGEOMINAS on the Red Cross, the police, and Defensa Civil offices hazard map and Defensa Civil on contingency in Armero had all been independently alerted plans, a videotape on natural disaster mitiga- to the concern of the Emergency Committee tion, a presentation by each mayor on local (Fig. 5). Local officials were therefore almost emergency plans, a summation of the well- certainly aware of the alert, but no formal regarded Libano plan and a forum for the decision to evacuate had been made. Before critical review and adjustment of the local 19:30 the abnormal conditions had apparently emergency plans. Following this discussion of ended, thus reducing concerns, and in Armero the agenda, a census of Los Nevados National residents heard repeated reassuring messages Park (that includes Ruiz) was presented, from Radio Armero and the village priest. restructuring of Emergency Committee sub- At 21:08 a strong eruption-produced seismic groups was debated and Cr. Perdomo made a signal occurred within Ruiz and seismographs presentation "praised for its clarity" of the were saturated for about an hour. A sequence of Defensa Civil contingency plan to be presented at least four andesitic to dacitic (58-65% SiO 2) at Libano. These details are included to pumiceous pyroclastic flows (one welded) and demonstrate that the Committee was not two surges were erupted and emplaced in rapid obsessed with the day's events, and indeed at succession (Calvache, 1986, 1990-companion 19:30, in response to an information request by volume; Janda et al., 1986, 1990-this volume; Red Cross radio, conditions were reported to Thouret et al., 1987), as was later interpreted have returned to normal (no ashfall) in north- from the 10-m-thick deposits on the summit ice ern Tolima. The meeting concluded with a cap and the thinner deposits extending several reminder of the important Libano meeting. kilometers from the crater (over 5 km down the By about 17:00 the minutes of the Ibagu~ Rio Azufrado); flow contacts show several meeting and interview comments suggest that meters of scouring relief. Magma eruption

Fig. 6. Aerial views of Nevado del Ruiz, before November 13, 1985, left; and after the eruption, right. The Azufrado Canyon borders the crater wall at top right. 170 B. VOIGHT temperatures exceeded 900°C (Katsui et al., and colluvium, lahars were mobilized in the 1986; W. Melson, oral commun., 1986). The channels of the major rivers. Farmers of the up- total dense-rock equivalent volume of eruption per valleys of the Lagunillas, Molinas, and products was 1.9 × 10 7 m 3 (Calvache, 1990- Guali reported hearing lahars as early as 21:15 companion volume). (Pierson et al., 1990-companion volume); initial The flows and surges melted such of the peak flows were into recession before the snowpack and carved a radial pattern of chutes climax of the subplinian eruption column at and channels in the 21-km 2 summit firn and 21:30 (Fig. 7). ice cap and outlet glaciers (Fig. 6; cf. Thouret, Bernardo Salazar and Rafael Gonzales, CHEC/ 1990-companion volume). Some meltwater Risk Comit~ staff, tending a seismograph 9 km sank into crevasses already choked with from the crater at E1 Arbolito had heard several pyroclastic debris, and large slabs of ice, firn strong explosions - one of which "lighted the and rock were pried loose from canyon walls raincloud over Ruiz lake a lamp" (SEAN, 1985), and avalanched into the headwaters of the Rio and Salazar radioed the Risk Comit~ in Azufrado and the Lagunillas (Pierson et al., Manizales (Fig. 5). Notified by the ComitY, the 1990-companion volume). Torrents of Governor of Caldas then called several commer- meltwater cascaded from the ice cap into the cial radio stations at various times between river channels flanking the cone. By entraining 21:30 and 22:30 to issue, in his own voice, a hot pyroclastic deposits, glacial drift, alluvium "red alert" for Caldas communities living

~ariquita

Fresnoo

He~veo

Sat~ndqa M

Murillo eLerida

Rio

PyroclasticDeposit

m ActualMud Flow i 10 M,ies i I 0 5i ,'0 15 K,Iorneters Ibagu61

Products of 13 November eruption. Contours show thickness of taphra deposit (airfall) [Comit~ data]. Compare with hazard map

Fig. 7. Products of November 13 eruption. Contours show thickness of tephra deposit (airfall) [Comit~ data]. Compare with hazard map. THE NEVADO DEL RUIZ VOLCANO CATASTROPHE: ANATOMY AND RETROSPECTION 171 along rivers from Ruiz (unpublished cor- before terminating in the Rio Magdalena (R. respondence from two Manizales radio stations, Janda, oral commun., 1986). Nov. 26, 1985). The information was passed A small quantity of muddy flow from the along to national emergency representatives Lagunillas would ultimately reach the and to national radio. Magdalena, but upstream matters were more The rate of tephra production peaked at about complicated. The Rio Lagunillas, about 50 km 22:00 and then gradually declined (Carey et al., long from its headwaters on the Nevado to its 1986). The pumice scorched many roofs but debouchment on the Armero plain, is fed by the caused no fatalities. Meanwhile, the lahars Rio Azufrado. Above the confluence, peak flow were racing down-valley, growing more on the Lagunillas was about 700 m 3 s -1. On voluminous by scraping a meter or so of rain- the 49-km-long Azufrado, however, peak flow soaked colluvium from valley walls, and en- cross-sectional area typically ranged from 2200 training several meters of bogs, sediment and to 2600 m 2 (Mojica et al., 1986). With an pore-water from the valley bottom (Mojica et average flow rate of roughly 10 m s -1 (72 km al., 1986; Pierson et al., 1990-companion in 2 hours), discharge averaged about 25,000- volume). The mud flowed in surges, "because 30,000 m 3 s-l; and astonishing peak value of where there were bridges the flow would be 48,000 m 3 s -1 occurred about 10 km down- held back, and then as soon as it (the bridges) stream from the source (Pierson et al., 1990- busted, it started roaring down again (J. companion volume). For comparison, this is Lockwood, written communication of eye- roughly equivalent to the wave produced by the witness accounts). Stratigraphic observations evulsion of practically the entire reservoir over confirmed that the lahars were multi-pulsed the world's largest arch dam in the gigantic events which were sustained for more than an Vai6nt, Italy, landslide catastrophe of 1963. hour on the flanks of the volcano, and for more The Armero story is essentially the Azufrado than 90 minutes further downstream (R. Janda, story, with the headwaters of the Lagunillas oral commun., 1986). The initial lahar wave was playing a subdued role. One eyewitness got it the most dilute and had the highest peak stage. exactly right (J. Lockwood, written commun., Peak flow depth in all lahar-scoured channels 1986): "the Lagunillas was flowing rather except Nereidas exceeded 10 m. Peak flow diminished. It was the Azufrado that brought velocities, determined from deposit supereleva- everything down and finished the little town tion at bends, were generally 5-15 m s-1 (Jan- downstream." da et al., 1986; Rojas and Borrero, 1986; Lowe At 21:45 to 22:00, officials in Ibagu6, aware of et al., 1986; Naranjo et al., 1986; Pierson et al., the paroxysmal eruption, attempted to order 1990-companion volume), typical speeds for the evacuation of Armero (Fig. 5). Engulfed in lahars on stratovolcano flanks. Subsequent a torrential, ashy rainstorm, Armero experienc- flow pulses were concentrated with debris, ed power and communication difficulties. By transporting 2-m-wide blocks, with late pulses then the Azufrado lahar, 30-40 m high, had becoming smaller and more dilute. Lahars from rounded the Villahermosa bend and was but the Molinas and Nereidas gathered with a peak half an hour or so from its confluence with the flow rate about 13,000 m 3 s 1, entered the Rio Lagunillas. The noise of the passing lahar Chinchin~, then flowed more than 70 km to the made it necessary for near-channel residents to Rio Cauca. The village of Chinchin~ was struck shout to be heard: "it was a supremely horrify- soon after 22:30; a partial evacuation had taken ing thing; we thought that our time had place, but 1,100 died. Lahars of the Guali flow- come.., you couldn't talk to a person 50 cm ed 90 km with a peak rate about 20,000 m 3 away from you because you couldn't hear" s 1, thinning somewhat beyond Maraquita (eyewitness testimony: J. Lockwood, written 172 B. VOIGHT

commun., 1986). Strong vibration was felt hours - almost precisely recapitulating several kilometers from the river. Acosta's 1845 mudflow. A few fortunate sur- Roughly half an hour after Libano, Murillo vivors were carried on mud the full distance and Ambalema Defensa Civil radios warned from Armero to Guayabal. A smaller lahar Armero to evacuate (Fig. 5), the lahar broke the descended the Rio Lagunillas. The lahar that landslide dam at Cirpe and released a cool- inundated Armero lasted slightly more than 2 water flood wave that raced ahead of the hours, but each of its multiple surges lasted less luminous yellow hot lahar (eyewitness than an hour (Pierson et al., 1990-companion testimony: J. Lockwood, written commun., volume). A second major pulse struck Armero 1986). Time had run out for Armero. at 23:50, followed by half a dozen or more Disgorged at about 27,000 m 3 s -1 from the smaller surges. A final major pulse after 01:00 canyon at about 23:30 as a wave nearly 40 m lasted about 15 minutes (Pierson et al., 1990- high (Pierson et al., 1990-companion volume), companion volume). the muddy boulder-laden torrent divided into Many survivors took flight only after hearing three branches (Fig. 8). The largest branch commotion in the streets as the first flood broke from the Lagunillas channel and across waves struck the village. Electric power failed central and southern Armero with a flow depth and confusion reigned in the darkness. Though of 2-5 m and velocities to 8 m s -1, totally many attempted to escape on foot, over twenty- destroying the urban sector, and then generally one thousand died. Thousands of the injured following the Rio Viejo - an old Lagunillas managed to reach high ground, but by noon channel - passed Santuario to barely form a next day, only 65 of the one to two thousand fan in the Rio Magdalena. Another branch residents still trapped alive in the Armero mud deflected northward and carried a divide to had been rescued. Many experienced great dif- reach the Rio Sabandija, which it dammed for ficulty in extracting themselves from the

1985 ~1985 • UENOeZW,,,';'t t AA E L;:;,°L~o'L ,~ ..'9

Mudflowdeposits of 13 November1985 near Armero, compared with reconstructed mudflowsof 1845and 1595.[after Mojica et al.]

Fig. 8. Mudflow deposits of November 13, 1985 near Armero, compared with reconstructed mudflows of 1845 and 1595 (aider Mojica et al., 1986). THE NEVADO DEL RUlE VOLCANO CATASTROPHE: ANATOMY AND RETROSPECTION 173 viscous mud. the waters were dragging beds along, overturn- Eyewitness accounts give an insight into the ing cars, sweeping people away... (we) went lack of decisive action, as well as the flavor of back into the hotel, a three-story building with catastrophe (J. Lockwood, written commun., a terrace, built of cement and very stur- 1986): dy... suddenly I felt blows, and looking "We witnessed what happened in Armero... towards the rear of the hotel I saw (something) we were in the city.., that night at 7:00 in the like foam, coming down in the darkness ... it evening it was raining heavily. We took shelter was (a wall of) mud approaching the hotel, and in the hotel, had dinner, then the weather settl- sure enough it crashed against the rear of the ed down, it got hot, then a group of us went out hotel and started crushing walls.., and then to look for a bar to play billiards and when we the ceiling slab fractured, and.., the entire returned at 10:50 the ash started falling.., we building was destroyed and broken into pieces, woke the (geology) professor up and our col- since the building was made of cement, I leagues, and when the professor saw that the thought that it would resists, but it (the particles were of lapill size, he informed us that bouldery mud) was coming in such an over- we should pack because we were leav- whelming way, like a wall of tractors, razing ing... We didn't hear any kind of alarm, even the city, razing everything... (then) the when the ash was falling and we were in the university bus that was in a parking lot next to hotel.., we turned on a radio.., the mayor the hotel was higher than us (on a wave of was talking and he said not to worry, that it mud), on fire, and it exploded, so I covered my was a rain of ash, that they had not reported face thinking, this is where I die a horrible anything from the Nevado, tostay calm in our death.., there was a little girl who I thought houses. There was a local station and we were was decapitated, but what happened was that listening to it, when suddenly it went of the her head was buried in the mud... A lady told air.., the electric power went out, and that's me, 'look that girl moved a leg.' Then I moved when we started hearing the noise in the air, toward her and my legs sank into the mud, like something toppling, falling, and we didn't which was hot but not burning, and I started to hear anything else, no alarm. Half an hour get the little girl out, but when I saw her hair before we had been out on the street and was caught, that seemed to me the most unfair everything was very quiet because no thing in the world... " businesses were open ... ash had fallen in the "All night long ash was falling. At 4:30 or afternoon and they had been informed to stay 5:00, more or less, the noise increased again, home with damp handkerchiefs, that it would then we thought that another mudflow was be nothing more than a rain of ash; so there coming, and sure enough, towards the side you were no alarms before.., the priest from could see something shining that was mov- Armero had supposedly spoken on a ing... " loudspeaker around 6:00 and had said the same "... And it started to be light, and that's thing, that there was no need to leave Armero, when we lost control because we saw that horri- because the population was alarmed on account ble sea of mud, which was so gigantic.., there of the story about the dam.., when we went were people burried, calling out, calling for out the cars were swaying and running people help, and if one tried to go to them, one would down.., there was total darkness, the only sink into the mud.., so now you must start light provided by cars.., we were running and counting time as before Armero and after about to reach the corner when a river of water Armero... it's like living and being born came down the streets.., we turned around again.., when you analyze that you realize screaming, toward the hotel, because already that a world definitely exists that is stranger 174 B. VO1GHT than the normal world in which we live." were not adequately described as significantly Another witness reported "we were listening larger, met with scepticism from local to the radio, but at no time were we told to authorities and populations over the need to leave the town. They simply said to keep cool evacuate" (Tomblin, 1985b, 1988). Following and remain calm.., perhaps it was the mayor, the late-afternoon ashfall, Radio Armero and I'm not certain.., it was a local station, from the church public address system had advised Armero. When we were listening to the radio calm, and even by 23:00, the mayor of Armero, we lost the radio wave, and started to tune it in ham radio contact with Ibagu~, was suffi- because we thought the radio was broken, but ciently uncertain of the threat that his family 15 seconds later the lights went out, then we remained in the village. His last words reveal realized that what happened was that (the mud) his surprise: "Wait a minute, I think the town first destroyed the transmitter further is getting flooded!" (Time, Nov. 25; Tomblin upstream." 1985b). A national TV station also had broad- Another: "I talked to a young man from the cast news of the eruption, although many Red Cross who was caught by the mud and remember the message as advising no cause for almost killed, he was very badly wounded. He alarm (P. Bolton and D. Mileti, written com- said that further upstream there is the munication) - a message appropriate for dam.., the water that had been dammed pass- Manizales but not for Armero. In fact, Radio ed by, the mud came behind, and they realized Armero was playing cheerful music when that the volcano had exploded and went out to power failed and mud engulfed the station. The warn the people..." Similarly, an official inertia of the "be calm" advisories following discussing on Red Cross radio the need to the mid-afternoon eruption was apparently not launch a full-scale alarm reported, "the water overcome by subsequent events, and strongly is coming!", whereupon the radio went dead. influenced the crisis decisions of both officials Some survivors reported that friends or and the general population. relatives had called and urged them to leave. Their accounts suggest that no systematic, Aftermath specific evacuation orders were issued by officials, although in a few cases representatives "Evening was coming on, but the town, once so noisy of various agencies may have taken action as at this hour, was strangely still. The only sounds were some bugle calls echoing through the air, still golden individuals. Thus one survivor reported (J. with the end of daylight..." (Camus) Lockwood, written commun., 1986). "From Paturri they called at 10:00 at night and one of In the frantic and mournful days that follow- the firemen in Armero went out on the streets ed, the ranks of the Volcanic Risk Comit~ swell- blowing a whistle and setting off the alarm, ed with volunteer scientists from many nations. alerting the people, but the people didn't want The geologic events of November 13 had yet to to come out of their houses. They said it was a be understood, and more important, danger lie; because the priest from Armero, the priest still existed for perhaps 80,000 people. More that said the rosary in the afternoon said that than 90% of the ice cap remained on the Nevado nothing had happened, not to be alarmed." (Thouret et al., 1987, 1990-companion volume), Similarly, a limited evacuation initiated by and although partly armored by pyroclastic Red Cross in response to events of the afternoon debris, the potential for eruption-generated "two hours later appeared to be unnecessary mudflows remained high. To aid the ComitY, and then became unpleasant because of heavy the USGS immediately dispatched (with OFDA rain and nightfall. Consequently, the report of support) about ten veteran volcanologists and new explosions at 9:00 that evening, which mudflow specialists armed with sophisticated THE NEVADO DEL RUIZ VOLCANO CATASTROPHE: ANATOMY AND RETROSPECTION 175 instrumentation. A six-station telemetered October 1988 two additional stations were in- seismic network and a tiltmeter array were stalled on the Rio Azufrado and the Rio Recio. operational by the end of November (Zollweg et The station locations do not allow appreciable al., 1986; Banks et al., 1986, 1990-this volume). lead time for some village, and the need re- Because Ruiz was often obscured by clouds, the mains for near-source monitors on all major telemetered seismographs provided a first line rivers fed by Ruiz. of defense for recognizing the onset of new erup- On the evening of January 3, vigorous tive activity (Decker, 1986). seismicity began at 23:20, saturating seismo- In addition, the U.S. Department of Defense graphs within an hour (SEAN, 1985). Low- sent two C-130 aircraft and 12 helicopters from frequency tremor began to dominate the the U.S. Army Southern Command to assist in seismic records at 01:28 (January 4), accom- medical evacuation, rescue and hazard mis- panied by B-type earthquakes and explosion sions; four Sikorsky Blackhawks remained un- events. Initially, the summit was not visible, til mid-December to fly support missions for but ashfall was noted at 03:00. Evacuation of volcano monitoring. For these services, OFDA about 15,000 residents from valleys of major was billed by the Department of Defense for the rivers began at 06:00. It took more than 3 hours phenomenal amount of $2,124,000: DOD there- to make the necessary decisions and pass the by claimed 78% of the total assistance provided alert to the public, 4:32 hours if reckoned from by the U.S. government. Yet the need was real, the onset of low-frequency tremor, 6:40 hours since Ruiz reached above the elevation capabili- from the onset of strong seismicity - times far ty of other helicopters in service in Colombia. in excess of that required for a lahar to reach Underscoring this fact, a Colombian helictopter endangered communities. Fortunately, tephra vanished without a trace over Ruiz later that production was minor and no hazardous lahars month, and several crashes have occurred. were generated by this event; unfortunately, Recognizing the need for a second line of the circumstance was regarded as a false alarm defense to ensure that no lethal mudflow could and thereby influenced future hazard decisions occur without a public alert, Risk Comit~ chief (Fig. 9). Pablo Medina and R.J. Janda in December and Since November 13, worrisome signs of in- January tried to gain USGS/OFDA approval for a stability in the thin septum separating the critically-located 7-station lahar-warning net- crater from the Azufrado headwall had raised work with GOES satellite telemetry, technology the possibility of a catastrophic rockfall and that had been demonstrated at Mount. St. lateral blast. EDM-monitoring, however, prov- Helens. However, USGS included overhead and ed the wall to be stable and not an immediate placed a cost of $290,000 on the project, and the threat to the population below (Voight, 1986; proposal was rejected by OFDA on the grounds Voight et al., 1986, 1987). of cost and technical complexity. OFDA, in fact, By February 1986, the Risk Comit~ had was anxious to disengage from operations at evolved into a well-equipped Observatorio Ruiz. It remained for the Japanese Ministry of VulcanolSgico de Colombia (ovc) under IN- Construction, in February, to provide two GEOMINAS, with constructively phased-out sup- ground-telemetered, multiple-cable lahar detec- port by USGS and OFDA. tors. Detector location was based on accessibili- ty and line power, with receivers at OVC, Retrospection Manizales (Rio Molinos) and the police station at Villahermosa (Rio Guali) containing an "In this respect our townsfolk were like everybody else, wrapped up in themselves ... they disbelieved in audio alarm and paper printout indicating catastrophes. A catastrophe isn't a thing made to man's which cables have broken and at what time. In measure; therefore we tell ourselves that it is a mere 176 B. VOIGHT

COE = National Emergency Com~ OVC = Volcano Observatory (Volcanic Risk Comit6 before Feb 1986) DC = Defensa Civil Post 13 November 1985 Linkages (sedes system) ..... Suggested Improved Unkages (parallel system)

i_~ Governors | (Provinces) i Radio, TV i

, .~ (Towns)Mayors i I Public Radio, TV Receivers I ' I i

o Manager Public O3

Repeated ~/~,. Messages & i i t l Personal Emergency Radio I '/~- Radio Receivers ] EVACUATION DECISION AND COMMUNICATION CHAIN POST 13 NOVEMBER 1985 Fig. 9. Evacuation decision and communication chain in January 1986. Decision-making was not efficient for the minor eruption of January 3-4, 1986; fortunately, no mudflows reached populated areas.

bogy of the mind, a bad dream that will pass away. But ing Martinique in 1902. Fortunately, no major eruption it doesn't always pass away and, from one bad dream to took place; but unfortunately, volcanologists were blam- another, it is men who pass away ... because they ed for their 'cry wolf forecast." haven't taken their precautions." (Camus) Volcanic emergency management is faced Before attempting to consider the factors that with a seldom-win situation. What does one cry specifically contributed to catastrophe at Ruiz, out when there may be a wolf?. The volcano- let us take a step back to look over some pro- logist's problem is to record accurately the problems of volcanic-hazard management. One babilistic chance of error in event forecasting predicament illustrated by Decker and Decker and hazard delineation, and yet maintain the (1981) mimics on a larger scale the January specificity and credibility necessary to en- 1986 event at Ruiz: courage appropriate government and management action, and public acceptance and "In 1976, a small eruption of La Soufri~re on response to the message. Guadaloupe led to a massive and expensive evacuation of 74,000 residents because of the threat that the erup- At the same time, predictive capability and tion might climax in catastrophic nudes ardentes risk assessment are only the first stage in the similar to those of the Mont Pel~e eruption in neighbor- complicated process of volcanic-hazard manage- THE NEVADO DEL RUIZ VOLCANO CATASTROPHE: ANATOMY AND RETROSPECTION 177 ment. To illustrate, three models are shown, all nosticate doom on a crude probabilistic basis, relatively simple but each successively more in- and the pragmatic decision makers. In the opi- volved. In the most elementary model, volcanic- nion of one experienced observer of the political risk assessments are directed toward govern- scene (Hess, 1987): ment authorities, who then have the option to develop policy and to create a management "Scientists are generally viewed with less regard than structure capable of responding to the perceiv- any other group on the Hill. From my observations, they are perceived as people who don't understand the ed hazard (Fig. 10a). Unfortunately, the win- system, who have little ability to talk with people out- dow of meaningful communication opens infre- side their profession, and who think that being correct quently between the scientists, who prog- is always enough... (Scientists) must learn that

Prehistoric, Historic Events Emergency / Lgs-.J-IMsnegsmsnt I Numbers are example probabilities

-X HISK L;ommunlcahon

C. Hypsrstrass

inlmlzed -- ~ sequence~

Emergency ~Appropriate The Public Extreme Management Event of Fortune L__oo

i ......

MODELS OF POLICY-MANAGEMENT-CONSEQUENCE RELATIONSHIPS

Fig. 10. Three models of policy-management-consequence relationships, each successively more involved. Events at Nevado del Ruiz progressed from the simplest model (a) in early 1985, to model (b) when the September 11 eruption served as a prime precursor event that closed the credibility gap, and finally to model (c) on November 13, 1985. 178 B VOIGHT

budgets, limited time frames, regulatory impacts, and influence policy opens momentarily for the risk public accountability are all factors that must be con- evaluators, and a management structure is sidered by Congress. Scientific facts are not the only spawned to deal with future events. criteria in public decision making." With the onset of a damaging and perhaps A result is that government may choose not to lethal event, existing management structure is employ emergency management. tested, as seen in the third model (Fig. 10c). The Yet an adequate emergency plan and critical schedule of decision making may be influenced risk levels must be established before an ade- by risk communication, but that mechanism is quate response can be expected to result from part of the management plan. Management is geophysical warnings. Further, if the public is "successful" if the consequences of the event to respond appropriately, they need to be con- are minimal, although the vicissitudes of for- vinced that the threat is real, they may need to tune may ultimately decide the outcome. believe that they personally are at risk, and Viewed in this light, the efforts at Ruiz in ear- they need to know precisely what to do and ly 1985 may be represented by the simplest where to go (Perry et al., 1981). Some of these model. Risk evaluations by UNDRO specialists elements were apparently missing at Mount St. found an audience - of sorts. But apart from Helens, when the public ignored restricted the preliminary disaster plan unveiled by zones and exerted pressure to have restrictions Defensa Civil in May, progress was sluggish removed, and when political interests influenc- and uncoordinated. INGEOMINAS did not yet ap- ed restricted-zone boundaries (Saarinen and pear fully committed to the task, in part owing Sell, 1985). to a lack of personnel experienced in volcanolo- An event which possesses the form, if not the gy, and probably also from a lack of conviction substance, of danger may therefore be a prere- that the threat was real. Thus (B. Martinelli, quisite for attracting the proper attention of written commun., 1986): "the directors of IN- government officials, of the population-at-large GEOMINAS always saw the situation of Nevado and even of hazard assessment organization. del Ruiz as a temporary emergency. Thus they Anxiety in sufficient quantity and properly openly opposed the creation of a permanent distributed is a proven catalyst to action. Thus observatory in Manizales, prefering to concen- the prime precursor event triggers action, not trate all monitoring activities in Bogot~." only among risk evaluators but also among Perhaps also lacking conviction, and possibly stakeholders, as seen in the second model. unwilling to act without OFDA funding, USGS Stakeholders are those who are exposed to administrators were willing to part with direct or indirect losses or mitigation costs from seismic transducers from its supply stock (for a hazardous event - the property owners, the which they were reimbursed) but were not non-risktaking cost bearers, financial institu- motivated to release specialists in volcanic tions, local, state and federal policy makers, emergency management or seismology. In any and the planners and insurers (Petak and event, with the perception of a vaccuum of Atkisson, 1982). Stakeholders wield political responsibility at the national level, it was clout. Just as a lightning discharge in the necessary for citizens and industry at a uprushing ash column creates an elec- regional level, within the shadow of the tromagnetic surge, so the prime precursor volcano, to gather a Comit~ to consider event closes, if only momentarily, the credibili- systematic monitoring and risk planning. ty gap separating stakeholders from risk Under such circumstances a rivalry between evaluators (Fig. 10b). With the gap closed, risk the Comit~ and INGEOMINAS seems understan- communication occurs and stakeholders are ac- dable. tivated. Pried by the stakeholders, a window to The volcanic hazard map requested by THE NEVADO DEL RUIZ VOLC.~NO CATASTROPHE: ANATOMY AND RETROSPECTION 179

Tomblin in March and urged by Hall in early alerted and were told to sound the alarm if May still had not been prepared by September, necessary. Information about the "end of abnor- and consequently management plans then in mal conditions" at 19:30 was also received, existence lacked the necessary focus toward although it certainly produced a signal con- areas targeted for high risk. The relative effi- tradictory to the succeeding alert. Messages ciency of UNESCO's early response had been concerning the evening magmatic eruption ar- checkmated, lacking a formal petition from the rived by various sources, including commercial Bogot~ government. radio beginning 21:30 to 22:30, which remained With the September 11 eruption of Ruiz as available for messages from regional or na- the prime precursor event, Colombia moved in- tional authorities, and Defensa Civil radio from to the realm of the second model. The gap of at least three localities after 22:30. River credibility closed, opening the doors to cen- observers' information probably influenced tralized power in Bogota. A National Emergen- evacuation suggestions to Armero from Libano, cy Committee emerged to coordinate activities, Murillo and Ambalema (El Tiempo, Dec. 4, INGEOMINAS became energized to take serious- 1985); thus the basic message of Defensa Civil ly and to complete the hazard map, and various had gotten through, at least indirectly, by regional committees accelerated their efforts. about 22:30. Red Cross radio from Armero to Responses to Colombian requests for assistance Ibagu~ was also connected, probably before were facilitated by enhanced credibility and the 23:00. Though the timing of many of the technical support essential to hazard manage- messages was probably too late to avert ment was provided by a flow of international casualties, numerous lives could have been sav- expertise. ed had prompt action been taken. By October, the management structure ap- It appears that the catastrophe cannot be peared to be relatively effective but when put to primarily attributed to technical breakdowns the test a month later - as in the third model in communications, which had succeeded in - the system failed. providing sufficient in imperfect information. Of the numerous factors that played a role in The primary problem was one of intent and the catastrophe, the following may be under- value; that is, the lack of will to act decisively. scored: (2) Response to an evacuation alarm issued (1) Power failures and communications break- immediately after the onset of the afternoon downs are common as a direct or indirect result phreatic eruption would have enabled no lives of volcanic eruptions; indeed they were an- to be lost, although by itself this event produced ticipated. Communications problems reported no long-reaching lahars, and thus could have at Ruiz included the failure of repeated late- provoked a charge of "false alarm". An alarm hour attempts to convey an evacuation message at the onset of the magmatic eruption would directly from Ibagu~ Defensa Civil to Armero also have provided enough time for a well- Defensa Civil and the failure of an upstream prepared Armero population to escape with lit- observer on the Lagunillas to directly contact tle loss - the two hours of lead time that had Defensa Civil in Armero by transmitting radio. been forecast by INGEOMINAS proved correct - However, review of technological communica- provided that the given alarm and public tions for November 13 reveals that the system response were swift, and the possibility of a contained many redundant elements and on the false alarm was accepted. Nevertheless, at whole worked successfully (Fig. 5). By 17:00 Chinching~ where the Governor of Caldas gave representatives of Defensa Civil, Red Cross and the alarm as early as 21:30, casualties were National Police, all fixtures in Armero emer- high and perhaps inevitable due to the realities gency management, had been independently of inefficient response and the much shorter 180 B. VOIGHT lead time. this level could be attributed to the delays in- (3) The scientists, emergency managers and herent in the entire chain of information flow. particularly the local officials therefore had The last links of the chain were at the village their backs to the wall. The risks were known level. With earlier essential links of the process early and well enough, but provincial and na- much delayed, time for critical review at the tional government made the conscious decision village level was minimized. Although the am- not to evacuate the villages unless and until the ple attention given to the evacuation possibility precise moment that danger could be in mid-September could have encouraged ade- guaranteed. To evacuate before the event, quate preparations and development of sound especially for an indefinite period, would have decision criteria, progress in Armero may have caused enormous problems and political risks been contradicted by government attempts to for a financially strapped government in pro- allay local fears of the Cirpe landslide dam. viding the funding, facilities and resources Also, detailed hazard mapping implied the lack needed to support a transported community. of any safe, easily reached refuge - thus any Security against looting in evacuated areas, decision to evacuate implied a major logistical and even the possibility of insurgents taking undertaking. In mid-October the mayor la- advantage of crisis conditions, could not be mented the state of frustration and confusion overlooked. And on the whole, the population after the hazard map had been released - and would have resented leaving their homes and noted a loss of "confidence in the veracity of in- having their livelihood and comfort disrupted; formation"; clearly he needed help. One may lacking a paroxysmal event, criticism surely surmise that some progress had been made in would have been severe. Therefore, fearing preparation for the November 15 Libano both the economic and political cost, no official meeting, but it was too late. was willing to take the responsibility for the (5) Preliminary evacuation plans for Armero order unless the damaging event - for Armero were drawn up by the local emergency commit- a long-reaching lahar, not merely an eruption tee, but apparently not many individuals had - could be predicted with almost absolute cer- yet seen them. "What-to-do" information had tainty. For science - whether the scientists ac- not yet been meaningfully communicated to the tually recognized it or not - it became more a population and the people were thus not ade- matter of detection than of prediction. quately prepared to act either individually or Thus, the authorities on the whole acted not with official instructions. unreasonably but were unwilling to bear the (6) The response of civil defense was initially economic or political costs of early evacuation hampered by the lack of specific hazard infor- or a false alarm. Science accurately foresaw the mation and maps, although Defensa Civil had a hazards but was insufficiently precise to render rather complete regional plan for volcanic-risk reliable warning of a devastating event at the management on paper by early May. Although last possible minute. Therefore, catastrophe progress was also restrained by the lack of had to be accepted as a calculated risk, and this cooperation between provinces facing similar combination - the limitations of predic- hazard, these plans were re-evaluated and tion/detection, the refusal to accept a false modified after September 11 and the process of alarm and the lack of will to act on the uncer- adjustment accelerated after October 7. Never- tain information available - provided the im- theless, it is one thing to have a plan on paper, mediate and most obvious cause of catastrophe. and another matter to implement it. Although (4)The local authorities were poorly pre- the need was clearly recognized, local plans had pared, funded and equipped to respond to an not yet been critically reviewed nor proof-tested emergency. In some respects, deficiencies at by warning and evacuation drills to identify THE NEVADO DEL RUIZ VOLCANO CATASTROPHE: ANATOMY AND RETROSPECTION 181 and correct weak links in the system. luck" - possibly impeded more decisive action (7) Preparation of hazard maps was not time- by authorities in Armero and created conflic- ly. Under ideal emergency-management cir- ting media advisories that retarded public cumstances the month-long dead time might response to sporadic evacuation suggestions. have been adequate; but at Ruiz the delay in Similarly, darkness and storm made conditions producing the hazards maps impeded the more difficult. mitigation effort, particularly for sites such as (10) What of the role of technology in the Armero and Chinchin~ where detailed local catastrophe? Would telemetered seismographs maps were essential for the completion and and real-time analysis have prevented refinement of local management plans and for disaster? The seismic records indicate that the educating local decision makers and the 15:06 phreatic eruption of November 13 was population on the specific nature of the lahar preceded by an energy release only slightly threat. In addition, other agencies' concern larger than background level, in contrast to the over the map as an obvious missing link clear premonitory increase prior to September diverted attention from additional points of 11. Following the November phreatic eruption, weakness in the mitigation chain. When releas- both earthquake activity and harmonic tremor ed in October, the preliminary map and ex- amplitudes increased abruptly, and tremor fre- planation were poorly distributed and the revis- quency decreased, suggesting to some experts ed map was not released until November 14, that a very short-term warning "might have the day after the catastrophe. been possible" for the 21:08 eruption; further, (8) Key individuals with specialized know- "cumulative tremor energies may have been a ledge can bring a hazard-assessment group better short-term predictor (i.e. a few days) than rapidly up-to-speed regarding key outcrops, earthquakes for the November 13 eruption, observations and local conditions, and can since an episode of low-level tremor began on significanctly reduce the time required for November 10 and continued until the eruption preliminary hazard evaluation and effective onset" (J. Zollweg and D. Harlow, written com- monitoring. It is vital that such individuals - mun., 1986; cf. Gil et al., 1987; Mufioz et al., whether or not skilled in volcanic emergency 1987). But Zollweg (written commun., 1988) has management - enter the hazard-evaluation since concluded after comprehensive examina- picture as early as possible; but non- tion of the records that neither cumulative Colombians most experienced in Ruiz geology earthquake energies, tremor energies, nor low- were not available until, in one case, after frequency earthquake or tremor patterns would September 20, for the second individual Oc- have been recognized as short-term precursors tober 16 and for a third only after November 13. prior to November 13: "The depressing conclu- In part this may reflect bureaucratic slug- sion is that all short-term precursors of seismic gishness or the reluctance of administrators in nature were ambiguous ... Your conclusion is such agencies as USGS and OFDA to release per- correct regarding the lack of seismic contribu- sonnel or provide pre-disaster support for tion to the catastrophe ... in this case there foreign assignments. In any case, this factor was to be virtually no recognisable short-term contributed to the delays in production of the warning." Ruiz hazard maps and in effective monitoring. However, to partially offset the lack of (9) Proclamation of the "end of abnormal con- telemetry, a seismometer 9 km from the crater ditions" after 19:30 caused confusion and skep- was manned and had direct communication ticism, inasmuch as it conflicted with alerts with Manizales (as recommended by the Italian concerning the evening magmatic eruption. team), and indeed the eruptive event was This factor - which might be classified as "bad detected and reported with time sufficient to 182 B. VO1GHT save most lives if appropriate action had follow- this respect, Volc~n Ruiz offers a better type ex- ed. The deficiencies of the seismograph net- ample of current hazard management than work, however real, did not contribute to the Mount St. Helens, for lessons derived from Ruiz November catastrophe. are more easily exported to other dangerous And what of the lahar detectors? These were volcanic regions, the majority of which are not an off-the-shelf item, and - perhaps sur- located in developing countries. prisingly - their utility was not widely In the 20th century at Armero, twenty thou- recognized until Armero was destroyed. Even sand - about 70% of the population - died in then they were rejected by OFDA. While the re- a virtual repetition of an event which killed quired lead time for assembly, installation and about one thousand - roughly 70% of the resolution of problems in the development and population - in the 19th century. In com- maintenance of communication linkages would parison, at Mayon volcano in the Philippines, have provided a hurdle to the effective use of where 1,200 died by surge and mudflow in the detectors before November 13, since then it has 1800s, the current population in the zone of been abundantly clear that detectors are essen- high risk approaches 800,000. Population tial for ensuring that no damaging lahar can oc- growth demands that people continue to live in cur without an adequate public alert. The cost areas of high hazard, many of which are places seems small when viewed as insurance. The of considerable agricultural and economic detector system installed by the Japanese cur- value, and as a result, natural catastrophes of rently plays an important role, although near- the future can be expected to be of staggering source monitors are still lacking to provide suf- proportions. It is therefore imperative that the ficient lead time for all communities at risk. developed countries pool financial and technological resources to establish programs that em- Lessons from Armero phasize pre-disaster preparedness in order to cope with such expected extreme-magnitude "Apparently it came to this: we might try to explain crises: the phenomenon ... but, above all, should try to learn (1) At Ruiz, it was not so much the impreci- what it had to teach us..." (Camus) sion of science, or inept last-minute decision- When Andr~ Coyne's magnificent but ill- making, or breakdowns of a few communication fated dam in southern France failed in 1954 systems at the vital moment - the failure was and the resulting flood erased much of the town to wait until the last possible minute. One can- of Frejus, the limit of the state-of-the-art design not expect emergency management to operate of arch dams was reached. An important lesson efficiently at that time scale; but this is often that emerged from this disaster was that what human nature seems to demand. It seems drainage is necessary to reduce hydrostatic less a matter of excessive confidence or false pressure downstream of thin arch dams: the sense of security than the lack of will to act in state-of-the-art of dam engineering was thereby the face of uncertainty and the unwillingness to advanced. Are there similar lessons to be gain- accept the costs implied by the finite probabili- ed from Armero? Or are there no new lessons to ty of a false alarm. Given that natural human advance volcanic emergency management - tendency, the developed nations must, on the merely and underscoring or reinforcement of one hand, continue to improve event prediction, lessons already compensated by past cata- event detection and communications technolo- strophes? gy for early warning, and urge drills with the Hindsight, while of no value to the former population at risk; but at the same time, they residents of Armero, can be of benefit if it helps must also seek to improved education in facing to avoid similar occurrences in the future. In uncertainty and false alarms, and seek to THE NEVADO DEL RUIZ VOLCANO CATASTROPHE: ANATOMY AND RETROSPECTION 183 achieve improve education in facing uncertain- convey respect for emergency measures and a ty and false alarms, and seek to achieve im- specific sense of what-to-do for individuals who proved understanding of policy science, so that may lack the capacity to read and write. When rationally conceived public policies are brought given, the warning must convince the public to to bear on these difficult problems in order that act. For example, evacuation research shows the expectations of crisis management might be that sirens, even when preceded by public reduced to a reasonable level. education, are insufficient to persuade people to (2) In moments of crisis, complex decision- leave their homes; sirens must be backed-up by making processes that rely on a chain of com- repeated verbal (possibly tape-recorded) warn- mand, hours of committee discussion and the ings from a respected public official (Perry et assumption of unstressed communication link- al., 1981; Mileti et al., 1986). ages, are neither efficient nor effective. The (4) Radio-communication problems are decision-making process can be much simpli- endemic in the steep canyons and dissected ter- fied by advance consideration of decision rains of volcanic country, and these are exacer- criteria. In countries with a highly centralized bated by the meteorological conditions which government, there is a tendency for officials for accompany eruptions. Communications tech- fear taking the initiative without higher nology for crisis response needs thorough authorization; consequently, a village leader evaluation. Satellite telemetry, microwave may wait too long to initiate the emergency ac- telephone systems standby power sources, and tion required to stave-off disaster. A document- redundancy all have a place in volcano ed poviso authorizing provincial and local ac- emergency management. tion in time of known volcanic crisis might be Yet electrical systems are susceptible to a feasible mechanism to surmount this difficul- lightning damage, and require trained tech- ty. nicians and the availability of electronic parts (3) The communication chain is no stronger to keep them functional. Consequently, crude that its weakest link: in emergency planning of traditional communications systems in current this sort, it is often the final links that are the use should not be totally abandoned and replac- weakest but the most important, yet problems ed by their high-technology counterparts; they in local emergency response are reported in should be employed as a redundant system. disaster case histories time after time. Manage- With an appropriate population distribution, ment's efforts tend to concentrate on the earlier for example, a linkage of individuals drumming links in the planning process, with less vigor on hollow logs, passing the message from one and critical review applied as succeeding links community to the next - "head for high ground and more local problems are encountered. In NOW" - may be more effective in an electrical the future, specific attention should be given to ashstorm at dead of night than a plan assuming local problems at an appropriate early stage. an orderly response to a radioed alert. Parallel rather than series communication (5) Hazard maps play a crucial role, yet one linkages may offer time-saving advantages might question whether the precise map boun- (Fig. 9). dary, or the precise contour defined in these The final link, that of providing timely, sim- maps, should make much difference to ple and effective information to the public can preliminary planning. In the future, when a be aided by news media, but they have neither volcanic-hazard specialist is called to a crisis the exclusive nor the prime responsibility. Civil situation in a country lacking such expertise, authorities must decide upon and implement part of his mission should be devoted to the direct and convincing measures to prepare and rapid production of a preliminary hazard map, alert the population. This information must however crude, from the materials available to 184 B. VOIGHT him. Whereas he may advise the host country's writing, the VDAP cache still lacked an off-the- geological bureay of its obligation to conduct in- shelf lahar-monitoring capability. Other pro- vestigations needed to improve and field-test blems have surfaced because of bureaucratic in- the map, the first priority is to produce a timely ertia. Support agencies still do not inevitably document-in-hand that will allow mitigation appreciate the necessity for baseline observa- planning to proceed without further delay. tional studies; these may be castigated as Clearly, individuals known to possess special "research" and too often are not considered as knowledge of the volcano should be sought out justified under pre-crisis conditions. Thus cur- immediately and included in early efforts to rent efforts are helpful, but they are inadequate evaluate hazard and produce maps. to address the problem on a global scale (Till- But at the same time, the utility of hazard ing, 1989). There is an urgent need to develop, maps should not be overemphasized. The pro- strengthen and expand such long-term production and media publication of a hazard map grams until developing nations acquire self suf- is not synonymous with risk communication. ficiency in volcanology. Indeed, the map itself, such a familiar tool to a (7) Finally, it is significant that such a small scientist, planner or engineer, may be virtually eruption at Ruiz was able to cause such a incomprehensible to individuals unfamiliar tremendous loss of life. Less than 5 x 106 m 3 of with the abstraction of reality on a sheet of magma ejected as pyroclastic flows was able to paper. Conventional education is no guarantee generate about 6 x 107 m 3 of lahar deposits of the ability to comprehend a map, a problem from 1 to 2 x 107 m 3 melt water (Calvache, that becomes more acute as one moves down 1990-companion volume; Pierson et al., 1990- the communication chain, since the literacy of companion volume). It had been known that individuals at risk cannot be assumed. such events could occur from eruptions on (6) The response at Ruiz underscores the need snow-covered volcanoes; at Mount St. Helens in for risk assessment, baseline monitoring data, 1982, for example, a blast of hot pumice and gas reliable prediction and detection technologies, caused a lahar with peak discharge about and the development and proof testing of alarm 14,000 m 3 s 1 (Waitt et al., 1983), and in South systems and evacuation plans in advance of America a spectacular and well-publicized volcanic crisis. Inasmuch as the majority of the lahar was produced by eruption-triggered world's high-risk volcanoes are in developing snowmelt at Cotopaxi volcano, Ecuador, in nations, scientific and economic support must 1877 (Wolf, 1878). Nevertheless, Ruiz provides be derived largely from developed nations and a particularly poignant reminder of the special international organizations. Most existing pro- sensitivity of the snow-capped volcanoes to grams for so-called rapid crisis assistance are even minor volcanic events. ad hoc in nature; they are inefficient and not truly rapid. A modest step in the right direction Epilogue is the Volcanic Early-Warning Disaster Assistance Program (VDAP), recently developed "Summoned to give evidence.., he has exercides the by USGS and OFDA to focus on South and Cen- restraint that behooves a conscientiouswitness. All the tral America (Banks, 1986). It includes such same, following the dictates of his heart, he had desirable features as development of scientific- deliberately taken the victim's side..." (Camus) political liaisons to avoid administrative At Ruiz, most of the elements considered hurdles and to facilitate rapid entry in time of highly desirable for successful volcanic emer- crisis, an experienced technical core team gency management were in place. The surficial capable of rapid response and a state-of-the-art stratigraphy surrounding the cone had been in- equipment cache. However at the time of vestigated to USGS "Cascade Range" stan- THE NEVADO DEL RUIZ VOLCANO CATASTROPHE:ANATOMY AND RETROSPECTION 185 dards, a geochronology had been established plague-stricken people, so that some memorial of the in- and post-glacial pyroclastic flows and lahars justice and outrage done them might endure; and to state quite simply what we learn in a time of had been mapped. Accurate and perspicacious catastrophe: that there are more things to admire in reports of key historical events were on record. men than to despise." (Camus) Advice was available through able and experienced foreign specialists in volcanic emer- Acknowledgements gency management, and monitoring equipment was provided. Many Colombians worked ad- In Colombia I worked as a member of the mirably and knowledgeably, and the conse- Comit~ de Estudios Volcanol6gicos, Manizales, quences of a potential eruption were adequate- with suport of OFDA and coordination by USGS. ly foreseen by risk evaluators. A hazard map Although drawing heavily on the work of the was produced over a month before the crucial ComitY, I accept full blame for the interpreta- event, and the problem had registered signifi- tions and opinions expressed here. If there is cant concern among national, provincial and value to this report it is due to the generosity of local governments, despite inevitable credibili- the numerous individuals who provided critical ty issues that arose from time to time. The but unpublished documentation, newspaper ar- magmatic eruption turned out to be small, and ticles and personal insights, and who were also its effects were not unprecedented - in fact the willing to correct erroneous or misleading historical data provided extremely close statements in earlier drafts. I therefore par- analogies. Unlike Coyne's dam in southern ticularly thank the following for contributions France, no startling issues arose to shift the of information, manuscript review, or photo- course of events. An yet in the the end, time pro- graphs: P. Bolton, M.L. Calvache, H. Cepeda, R. ved to be a luxury squandered. Decker, T. Ertekin, J.C. Griffiths, M. Hall, D. The catastrophe was not produced by tech- Harlow, R.J. Janda, B. Martinelli, H.J. Meyer, nological ineffectiveness or defectiveness, nor D. Mileti, E. Parra, T. Pierson, B.A. Shattuck, by an overwhelming eruption of unprecedented J.C. Thouret, J. Tomblin, E.G. Williams, S. character, nor by an improbable run of bad Williams, K. Young, and J. Zollweg. I also owe luck. Armero was caused, purely and simply, by a great deal to Jefe P. Medina J., and the entire cumulative human error - by misjudgement, Comit~ staff; OVTR Andinistas M.L. Calvache indecision and bureaucratic shortsightedness. and O. Ospina H.; F. Zambrano O. and A. In the end, the authorities were unwilling to Nunez, INGEOMINAS; N. Banks and the USGS bear the economic or political costs of early crew; acting U.S. Ambassador J.P. Osterling; evacuation or a false alarm, and they delayed and A. Swan and staff, OFDA. Translations action to the last possible minute. Catastrophe were provided by J. Lockwood, C. Faria, A. was the calculated risk, and nature cast the die. Lacazette, R. Holcomb, and particularly C. And so the lessons from Armero are not new Moy. Finally, but hardly least, I wish to express lessons; they are old lessons forged in human my gratitude to Judith Kiusalaas for her en- behavior that once again require the force of thusiastic, skillfull and unflagging efforts to catastrophe to drive them home. Armero could disguise my deficiencies in logical thought and have produced no victims, and therein dwells muddled attempts to properly render the its immense tragedy. English language. Some material in this paper was presented at the 1986 Lahar and Landslide "And it was in the midst of shouts rolling against the Workshop in the Philippines, and Institutes for terrace wall in massive waves that waxed in volume and duration, while cataracts of colored fire fell thicker Disaster Prevention in Tsukuba and Shinjo, through the darkness . .. that [it was] resolved to com- Japan. A preliminary version was delivered to pile this chronicle... [to] bear witness in favor of these the 1988 Ruiz Volcano Workshop in Manizales. 186 B VOIGHT

References J.L., 1986. The November 13, 1985 eruption of Nevado del Ruiz: Tephra fall deposit and column Acosta, J., 1846. Relation de l'~ruption boueuse sor- height modelling (abstr.). EOS, Trans. Am. tie du volcan de Ruiz et de la catastrophe de Geophys. Union, 67 (16): 403. Lagunilla dans la l~publique de Nouvelle Cepeda, H., Michael, J., Murcia, A., Parra, E., Grenade. C. R. Acad. Sci., 22: 709-710. Salinas, R. and Vergara, H., 1985a. Mapa Acosta, J., 1850. Sur les montagnes de Ruiz et de Preliminar de Riesgos Volc~nicos Potenciales del Tolima (Nouvelle Granada) et les ~ruptions Volc~n Nevado del Ruiz. Esc. 1:50.000. Memoria boueuses de la Magdalena. (Lettres a M. Elie de Explicativa, INGEOMINAS, Medellin, Oct. 7, Beaumont). Bull. Soc. Geol. Fr., ser. II, 9: 1985, 22 pp. 489-496. Cepada, H., Michael, J., Murcia, A., Parra, E., Banks, N., 1986. Hazard mitigation and rapid Salinas, R. and Vergara, H., 1985b. Mapa de response to crises with Mobile Volcanic Obser- Riesgos Volcanicos Potenciales del Nevado del vatories (abstr.). EOS. Trans. Am. Geophys. Ruiz. Esc. 1:100.000. Memoria Explicativa, IN- Union, 67 (16): 398. GEOMINAS Medellin, Nov. 12, 1985, 28 pp. Banks, N.G., Van der Laat, R., Carvajal, C. and Ser- Cepeda, H., Michael, J., Murcia, A., Parra, E., rano, T., 1986. Deformation monitoring at Salinas, R. and Vergara, H., 1986. Volcano risk Nevado del Ruiz, Colombia (abstr.). EOS, Trans. potential maps of Nevado del Ruiz, Volcano, Col- Am. Geophys. Union, 67 (16): 403. ombia (abstr.). EOS, Trans. Am. Geophys. Union, Banks, N.G., Carvajal, C., Mora, and Tryggvason, E., 67 (16): 407. 1990. Deformation monitoring at Nevado del CHEC, 1983. Investigacidn Geoth~rmica, Macizo Ruiz, Colombia: October 1985-March 1988. In: Volc~nico del Ruiz, Fase II Etapa A, Vol. I. In- S.N. Williams (Editor), Nevado del Ruiz Volcano, forme final, Bogotfi 119 pp. Colombia, I. J. Volcanol. Geotherm. Res., 41: Cu611ar, J. and Solano, A., 1985. Informe Preliminar 269-296. de las actividades desarrolladas. INGEOMINAS, Barberi, F., Martini, M. and Rosi, M., 1985. Informe Bogota, October 1985, 19 pp. de la misiSn tecnica para la evaluacion del riesgo Decker, R.W., 1986. Ruiz Volcano, Colombia: Trip volcanico del Nevado del Ruiz-Cordillera Central- Report, February-March 1986. U.S. Geological Colombia. Manizales 22 Octubre 1985. Ministero Survey, Menlo Park, 45 pp. (unpublished). degli Affari Esteri-Italia-Dipartimento Coopera- Decker, R.W. and Decker, B., 1981. Volcanoes. zione per lo Sviluppo, 18 pp. Freeman, San Francisco, CA, 244 pp. Barberi, F., Martini M. and Rosi, M., 1986. The Espinosa, A., 1986. Notas sobre la Actividad Italian Mission to Nevado del Ruiz, Colombia (Oc- Hist6rica del Volcan Nevado del Ruiz. IN- tober 1985): results, recommendations and some GEOMINAS, Cali, 17 pp. reflections. Per. Mineral., 55: 1-4. Gil C., F., Meyer, H.-J., Chouet, B. and Harlow, D., Barberi, F., Martini, M. and Rosi, M., 1990. Nevado 1987. Observations of long-period events and del Ruiz volvano (Colombia): pre-eruption obser- tremors at Nevado del Ruiz Volcano 1985-1986. vations and the November 13, 1985 catastrophic Hawaii Symposium on How Volcanoes Work. event. In: S.N. Williams (Editor), Nevado del Ruiz Abstr. Vol., Hawaiian Volcano Observatory, Hilo, Volcano, Colombia, II. J. Volcanol. Geotherm. Hawaii, Jan. 19-25, p. 90. Res., 42: 1-12. Hall, M.L., 1985 (mission report). Activities and con- Calvache V., M.L., 1986. The eruption of the Nevado clusions of visit by Dr. M. Hall to Volcfin Nevado del Ruiz volcano, Colombia (abst.). EOS. Trans. del Ruiz, Colombia. Instituto Geofisico, Escuela Am. Geophys. Union, 67 (16): 405. Polit6cnica Nacional, Quito, 4 pp. (English and Calvache V., M.L., 1990. Pyroclastic deposits of the Spanish versions with subtle differences). November 13, 1985 eruption of Nevado del Ruiz Hall, M.L., 1990. Chronology of the principal scien- volcano, Colombia. In: S.N. Williams (Editor), tific and governmental actions leading up to the Nevado del Ruiz Volcano, Colombia, I. J. November 13, 1985 eruption of Nevado del Ruiz, Volcanol. Geotherm. Res., 41: 67-78. Colombia. In: S.N. Williams (Editor), Nevado del Calvache V., M.L., Duque, G., Garcia, N., Salazar, Ruiz Volcano, Colombia, II. J. Volcanol. J.E. and Salazar, B., 1985. Memorias del Geotherm. Res., 42: 101-115. Seminario Sobre Riesgo Volc~nico del Ruiz. Nat. Harlow, D.H., Munoz, C. and Cuellar, J., 1986. Univ., Manizales. Rev. Vias y Transport., 12 (53) Seismic activity preceding the 13 November 1985 114 pp. eruption of Ruiz volcano, Colombia (abstr.). EOS, Carey, S., Sigurdsson, H., Fritz, W. and Naranjo, Trans. Am. Geophys. Union, 67 (16): 403. THE NEVADO DEL RU1Z VOLCANO CATASTROPHE: ANATOMY AND RETROSPECTION 187

Herd, G., 1982. Glacial and volcanic geology of the Jan. 19-25, p. 185. Ruiz-Tolima complex, Cordillera Central, Colom- Mufioz, F. and Meyer, H.-J., 1987. Seismic activity bia. Publicaciones Geologicas Especiales del IN- preceding the 13 November 1985 eruption at GEOMINAS No. 8, Bogota, 48 pp. Nevado del Ruiz. Hawaii Symposium on How Herd, G. and Comit~ de Estudios VulcanolSgicos, Volcanoes Work. Abstr. Vol., Hawaiian Volcano 1986. The 1985 Ruiz volcano disaster. EOS, Observatory, Hillo, Hawaii, Jan. 19-25. Trans. Am. Geophys. Union, 67 (19): 457-460. Naranjo, J.L., Sigurdsson, H., Carey, S.N. and Fritz, Hermelin, M., Velasquez, A. and Bustamente, M., W., 1986. Eruption of Nevado del Ruiz volcano, 1986. Aspectos historicos de eruptiones en el Colombia, on 13 November 1985: Tephra fall and Volcan Nevado del Ruiz. Bol. Vias, 13 (56), 37 pp. lahars. Science, 233: 961-963. Hess, J., 1987. Report from Washington. GSA News Parra, E. and Cepeda, H., 1990. Volcanic hazard and Information, Geol. Soc. Am., Boulder, maps of the Nevado del Ruiz volcano, Colombia. November, p. 248. In: S.N. Williams (Editor), Nevado del Ruiz Janda, R.J., Banks, N.G., Pierson, T.C., Calvache V., Volcano, Colombia, II. J. Volcanol. Geotherm. M.L. and Thouret, J.-C., 1986. Interaction be- Res., 42: 117-127. tween ice and pyroclastic sediments erupted dur- Perry, R.W., Lindell, M.K. and Green, M.R., 1981. ing the 13 November 1985 eruption of Nevado del Evacuation Planning in Emergency Manage- Ruiz, Colombia (abstr.). EOS, Trans. Am. ment. Lexington Books, Lexington, MA. Geophys. Union, 67 (16): 406. Petak, W.J. and Atkisson, A.A., 1982. Natural Katsui, Y., Takahashi, T., Egashira, S., Kawachi, S. Hazard Risk Assessment and Public Policy. and Watanabe, H., 1986. The 1985 eruption of Springer-Verlag, New York, NY. Nevado del Ruiz volcano, Colombia, and asso- Pierson, T.C., Janda, R.J., Thouret, J.-C. and Bor- ciated mudflow disaster. Report of Natural rero, C.A., 1990. Perturbation and melting of Disaster Scientific Research, no. B-60-7 (Japan snow and ice by the 13 November 1985 eruption Ministry of Education), pp. 61-91. of Nevado del Ruiz, Colombia, and consequent Lowe, D.R., Williams, S.N., Leigh, H., Connor, C.B., mobilization, flow and deposition of lahars. In: Gemmell, J.B. and Stoiber, R.E., 1986. Lahars in- S.N. Williams (Editor), Nevado del Ruiz Volcano, itiated by the 13 November 1985 eruption of Colombia, I. J. Volcanol. Geotherm. Res., 41: Nevado del Ruiz, Colombia. Nature, 324: 51-53. 17 -68. Medina J., P., 1986. The organization of the group to Rojas, R. and Borrero, C.A., 1986. Velocidad, monitor volcano Nevado del Ruiz in Colombia descarga maxima y volumen del lahar del Rio (abstr.). EOS, Trans. Am. Geophys. Union, 67 Chinchina, 13 Noviembre 1985. In: Proc. Sim- (16): 403. posio Internacional sobre Neotectonica y Riesgos Mileti, D.S., Bolton, P.A., Fernandez, G. and Updike, Volcanicos, Bogota, Colombia, Diciembre 1-7, R.G., 1986. Recommendations for improving the 1986. existing warning system for an impending erupt Saarinen, T.F. and Sell, J.L., 1985. Warning and tion of the Nevado del Ruiz volcano, Colombia, Response to the Mount St. Helen's Eruption. S.A.: and Advance Report. Committee on Nat. State Univ. New York Press, Albany, NY, 240 pp. Disasters, Nat. Acad. Pres, 10 pp. SEAN, 1985, Ruiz volcano, Colombia: Smithsonian Miller, C.D., Mullineaux, D.R. and Hall, M.L., 1978. Institution, SEAN Bull., 10 (11), p. 2-5. Reconnaissance Map of Potential Volcanic SimSn, P., 1625. Noticias Historiales de las Con- Hazards from Cotopaxi Volcano, Ecuador. Scale quistas de Tierra Firme en las Indias Occiden- 1:100,000. U.S. Geol. Surv., Misc. Invest., Map 1- tales. Casa Editorial de Medardo Rios, Bogot~i, 1072. 376 pp. (1892). Mojica, J., Colmenares, F., Villarroel, C., Macia, C. Thouret, J.-C., 1990. Effects of the November 13, and Moreno, M., 1986. Caracteristicas del flujo de 1985 eruption on the snowpack and ice cap of lodo ocurrido el 13 de Noviembre de 1985 en el Nevado del Ruiz volcano, Colombia. In: S.N. valle de Armero (Tolima, Colombia): Historia y Williams (Editor), Nevado del Ruiz Volcano, Col- comentarios de los flujos de 1595 y 1845. Geol. ombia, I. J. Volcanol. Geotherm. Res., 41: 177-201. Colomb., 14: 107-140. Thouret, J.C. and Calvache V., M.L., 1986. The con- Mufioz, F., Meyer, H.-J., and Harlow, D., 1987. sequences of the 13 November 1985 eruption on Seismic activity preceding the 13 November 1985 the Nevado E1 Ruiz ice cap (abstr.). EOS, Trans. eruption of Ruiz volcano, Colombia. Hawaii Sym- Am. Geophys. Union, 67 (16): 404. posium on How volcanoes Work. Abstr. Vol., Thouret, J.C., Murcia, L.A., Salinas, R., Perignon, Hawaiian Volcano Observatory, Hilo, Hawaii, N.V. and Cantagrel, J.M., 1985. Cronoestrati- 188 B. VOIGHT

grafia mediante dataciones K-AR y C-14 de los Seismologico de Costa Rica, Univ. Nacional, volcanes Compuestos del Complejo Ruiz-Tolima y Heredia, Costa Rica, 23 pp. and 15 p. Anexo. aspectos volcanoestructurales del Nevado del Vel~squez, A., Hermelin, M. and Bustamente, M., Ruiz (Cordillera Central, Colombia). Mem. VI 1986. Actividad del Volcan Nevado del Ruiz en Congr. Latinoamericano de Geol., Bogot~i, Tom. 1, epocas historicas. Bol. Vias, 13 (59), 20 pp. pp. 385-454. Voight, B., 1986. Gravitational failure hazards, Thouret, J.C., Janda, R.J., Pierson, T.C., Calvache Nevado del Ruiz, Colombia-Mission from 30 V., M.L. and Cendrero, A., 1987. L'eruption du 13 January to 16 February 1986. Comit~ de Estudios Novembre 1985 au Nevado E1 Ruiz (Cordillere VulcanolSgicos, Manizales, 20 pp. Centrale, Colombie); interactions entre le dyna- Voight, B., Calvache V., M.L. and Ospina H., O., misme eruptif, la fusion glaciaire et la genese 1986. Is the Arenas crater at Nevado del Ruiz en- d'ecoulements volcano-glaciaires. C.R. Acad. Sci. dangered by gravitational slope collapse? (abstr.). ser. II, 305: 505-509. EOS, Trans. Am. Geophys. Union, 67(16): 406. Tilling, R.I., 1989. Volcanic hazards and their Voight, B., Calvache V., M.L. and Ospina Herrera, mitigation: progress and problems. Rev. Geo- O., 1987. High-altitude monitoring of rock mass phys., 27 (2): 237-269. stability near the summit of Volcano Nevado del Tomblin, J., 1985a. Report of a mission from 25 Ruiz, Colombia. Proc. 6th Cong. Int. Soc. for Rock Sept-2 Oct. 1985. UNDRO, Geneva, 9 October, 5 Mechanics, p. 275. pp. and 6 Annexes. Waitt, R.B., Jr., Pierson, T.C., McLeod, N.S., Janda, Tomblin, J., 1985b, Ruiz - A sleeping giant awakes. R.J., Voight, B. and Holcomb, R.T., 1983. UNDRO News, Office of the United Nations Eruption-triggered avalanche, flood, and lava at Disaster Relief Coordinator, Geneva, Nov./Dec., Mount St. Helens - effects of winter snowpack. 1985, 7 pp. Science, 221(4618): 1394-1396. Tomblin, J., 1988, UNDRO's role in responding the Wolf, T., 1878. Der Cotopaxi und seine letzte Erup- volcanic emergencies. UNDRO News, Office of tion am 26 Juni 1877. Neues Jahrb. Mineral. the United Nations Disaster Relief Coordinator, Geol. Paleontol., pp. 113-167. Geneva, Mm'ch/April 1988, p. 7-10. Zollweg, J.E., Harlow, D.H., White, R.A., Norris, UNDRO, 1986, Volcanic Emergency Management. R.D. and Guendel, F., 1986. Post-eruption United Nations, New York, NY, 86 pp. seismicity at Nevado del Ruiz, Colombia: Implica- Van der Laat, R., 1985. Informe de la mision tecnica tions of the reservoir system and continued activi- para la implementacion de estudios de deforma- ty (abstr.). EOS, Trans. Am. Geophys. Union, cion del Volcan Nevado del Ruiz, Colombia. Un- 67(16): 406. pub. Informe del Observatorio Vulcanol6gico y