April 25, 2001
Natural and human-caused disasters hit libraries no less often than any other institution, yet the impact of losing a library collection far outweighs the dollar value assigned to the damage. From the fires that ravaged the ancient library at Alexandria and the modern Los Angeles Central Library to the floods and tornadoes that have destroyed libraries in the Midwest, our cultural heritage suffers with each disaster. The great number of libraries in the U.S. and around the world can't insure that unique copies of books and journals won't disappear into memory with the next emergency. Despite the unpredictable nature of disasters, libraries can minimize their losses through careful planning; with each occurrence, libraries can learn from other's experiences and improve the chances of saving their own priceless collections.
The recent earthquake in Puget Sound illustrates the damage that can occur even when library buildings survive relatively unscathed. The University of Washington's library was the hardest hit, with bookshelves twisted or collapsed, and tens of thousands of volumes thrown to the floor. Covers were torn, pages ripped, bindings broken and more (We Need, 2001). While few if any volumes will be completely lost, the cost of restoring them will be tremendous and the University of Washington Foundation is desperately seeking donations and funding; in the meantime, a significant portion of the collection will be unavailable. On a smaller scale, this story was repeated across the southern Puget Sound area in public and school libraries.
Libraries in different regions face varying potential emergencies. On the eastern seaboard, hurricanes and floods are always possible. In 1999, Hurricane Floyd struck the coast, mainly damaging collections in Virginia and North Carolina (Rogers, et al., 1999). In the Carrollton (Va.) Public Library, flooding soaked a quarter of the collection-1600 volumes representing the majority of the reference section-which had to be discarded. Water level in some North Carolina libraries reached shoulder level and entire collections were destroyed.
Tornadoes strike the Midwest every year, sometimes plowing through libraries in their paths, as was the case in Fort Worth, Tex., where a storm hit the newly renovated Central Library, sending glass shards from broken windows and an atrium skylight all through the building, along with rain (Sullivan, 2000). Damage was estimated to be several thousand dollars after insurance. In this case, water damage to the books and the building was significant, but cleanup was costly in part because of the tedious work of cleaning pieces of glass from a large portion of the collection.
No matter where libraries are located, fires and water damage (from sprinklers) are a constant threat. Arson in the Los Angeles Central Public Library in April 1986 caused the largest library fire in U.S. history with hundreds of thousands of volumes burned, smoke damaged or wet (Manning, 1986; Follow-up, 1986). Almost 350 LA Fire Dept. personnel were involved in the seven and half hour fight to save the library, including sixty different firefighting companies, as well as four salvage companies and many others. Despite millions of dollars in estimated losses, immediate salvage operations saved the vast majority of the collection.
Not all disasters are as immediately dramatic or even natural, of course. A number of pests, such as Anobiidae (deathwatch beetle), book lice, silverfish and others, love paper and bookbinding glue. They present enormous problems in some libraries, such as Yale, where several species of beetles imported in a set of rare 16th-century European books infested over 37,000 volumes (Morris, 1986). The ravages of acid paper documented in Terry Sander's 1986 film "Slow Fires" destroys collections just as surely as a faster blaze, and is estimated to affect over 73 million volumes produced since the mid-nineteenth century. American libraries have been fortunate to escape war damage in the past century (although not necessarily the effects of civil unrest), but elsewhere in the world, wars small and large destroy libraries like the public library in Vinkovci, Croatia that was bombed in 1991 (Cveljo, 1999) and the entire National and University Library of Bosnia and Herzegovina in Sarajevo that was destroyed by Serbian artillery in 1992 after being shelled until it went up in flames (Disaster Roundup, 1992). Most of its three million volumes were lost and citizens who entered the building to rescue the most valuable material were fired upon.
The key to meeting library emergencies and disasters lies in a comprehensive emergency preparedness plan, especially for smaller institutions without extensive in-house facilities and conservation resources. Jan Lyall (1996), of the National Library of Australia's National Preservation Office, describes disaster plans as detailing "the procedures devised to prevent and prepare for disasters, and those proposed to respond to and recover from disasters when they occur. The responsibility for performing these tasks is allocated to various staff members who comprise 'the disaster team'."
Lyall notes that every plan has three phases: before, during and after. Each of these phases require separate, but interrelated, plans. The "before" phase involves preparedness and prevention measures such as emergency medical supplies, plastic sheeting, improved security, regular building maintenance, and training of staff to deal with likely situations. The "during" phase is concerned with responding to the emergency itself. In the "after" phase, recovery plans go into effect. Since all emergencies are unique, no plan can cover all possibilities, but certain common elements such as water damage can be covered in greater detail.
She also notes that plans should consider all areas likely affected by the disaster, including personnel (staff, users and visitors); collections and records (both the collections and the institution's vital records); and the facility (building and equipment).
She recommends a thorough risk analysis be preformed before developing procedures, with risks ranging from events with high probability/high effect (e.g. earthquake) to low probability/low effect (e.g. vacuum cleaner malfunction). She notes that probabilities and effects will vary with institution because of varying geography, building design, political environment, etc. The lower probability/lower effect events can usually be prevented as part of regular maintenance and upgrades; the goal is minimize risks and to move as many events out of the higher effect categories through preventive measures.
As institutions prepare new plans, they will usually have an existing plan that can be revised. Improvements and additional procedures may be needed to meet high probability disasters, although they may depend on availability of funds, facilities such as freezers, and trained personnel. Since disasters may have wide-ranging effects, specific staff members should be responsible for specific areas. Finally, coordination during and after an emergency is extremely important; which staff has responsibility for declaring an emergency and implementing the plan must be identified.
David Weber (1990) recommends that all staff be aware of emergency procedures and selected senior staff keep a copy at home, complete with key phone numbers, lists of commercial sources for supplies, and a list of emergency response team members. He also recommends that all staff receive some amount of regularly updated.
For large institutions, good emergency response may depend less on the plan itself and more on the presence of a highly professional staff, skilled preservationist, and availability of immediate resources. Walter Henry (personal communication, April 25, 2001), Collections Emergency Coordinator for Stanford University Libraries (SUL), says that a small institution such as a historical society must depend more on a written plan, often written with the help of consultants, because they will lack the necessary expertise and experience. For a large institution such as Stanford, the highest-level managers, in consultation with conservators, will meet as soon as possible after the emergency occurs, assess the situation and come up with a specific plan to meet the situation-which will never be exactly the scenario covered by the written plan.
Stanford's plan (Stanford, 2000) calls for several teams to respond to emergencies, including
Henry said that the plan calls for the C.E.M.T. to meet annually to familiarize themselves with procedures; in actuality, meeting annually is only necessary if the most recent disaster is five or more years in the past. Likewise, the C.E.R.T. doesn't really exist because the supervisors have moved within the organization too often to maintain a stable team; fortunately, most of the basic tasks don't require much training so much as management skills and common sense. A trained C.E.R.T. can be most useful if multiple libraries have problems or if staff needs to deal with non-standard storage units such as malfunctioning compact shelving. Rebuilding C.E.R.T. is high on Henry's priorities for the coming year.
Henry notes the plan is usually a portrait of the most recent disaster and in the event of emergency, useful mainly as reference with phone lists for key personnel, businesses to contact for supplies, technical information, lists of everyone who needs be involved, how to deal with the media, facilities, etc. He notes that even with the best intentions, most of the people will have only a glancing familiarity with the plan and that procedures are often hypothetical-something that is supposed to happen usually doesn't. A good example is the pair of trailers full of disaster supplies that are supposed to be hauled from the storage yard to the affected library during an emergency; unfortunately, the facilities staff has always been too occupied with the emergency themselves to bring the trailers out.
Lyall (1996) identifies lack of awareness of an emergency plan as the main reason such plans fail-they are useless if they gather dust on a shelf and no one knows what they contain. Staff must be aware of their own responsibilities. Identifying emergency response teams may be useless if no regular training is offered.
Other reasons for failure include lack of commitment to the plan-some preparations may be difficult, costly or time-consuming (such as regular drills), and so they are neglected. Staff responsible for developing procedures may not take time to document the necessary details (Wright, 1988). Another is the unpredictable nature of disasters; surprises will always be a part of any disaster and the scene must be thoroughly surveyed. Finally, the disaster may be a much greater magnitude than ever expected and the entire building or collection could be destroyed (Lyall, 1996).
When a plan fails, the impact will likely be much greater in terms of the organization, the collection or the facility-any of which will further harm the institution's ability to function. The Santa Clara County Library reported that good preparation before the 1989 Loma Prieta earthquake, including emergency preparedness classes and defining a clear vision of roles, gave employees the clear heads they needed to meet the situation (Weber, 1990). Even if real disasters occur frequently, regular drills are the best way to maintain staff awareness and familiarity with procedures. Lyall (1996) concludes that the very most important factor is for all staff, particularly senior staff, to remain committed to the plan.
Thorough planning saved the collections of the Huntington Library and Art Gallery in Pasadena, Calif., from the October 1985 fire that began in an elevator shaft (Roberts, et al., 1988). The conservation team reported that the fire department reached the building three minutes after receiving the alarm call and suppressed the fire within twelve minutes. The team noted that the fire department conducted regular drills in the building and were extremely well prepared; because of the fire department's quick action, only a few items were badly damaged or destroyed, with most of the damage limited to soot throughout the building. The library collection of 25,000 volumes had to be treated for soot damage, with each book requiring careful vacuuming. Fortunately, the layer of dust on the books prevented the penetration of carbon and oily soot and real damage. The cleanup took approximately 500 hours.
Lyall (1996) notes that disaster plans' success can be measured in how they reduce the frequency of disasters, quantity and value of material lost and quantity requiring conservation, time required for a library to return to normal operations, time to repair the building, and how well staff can operate under emergency conditions. Henry (personal communication, April 25, 2001) argues that such indicators may measure the professionalism and expertise of managers and conservators as much as the plan itself.
Even the best plan at any institution can't guarantee success. Nonetheless, Henry notes plans are useful in that they help people feel better and calmer. Furthermore, in the event that all the experienced management and staff is away, the plan can guide the remaining staff until they return.
The Stanford University Libraries, a large academic library system with many branches and a century long history, provides a number of good examples of disaster response. Throughout the institution's century of existence, it has been afflicted at various times by earthquakes, floods, as well as insect infestations, vandalism, and theft. Two events in particular have been well documented and studied: the 1989 Loma Prieta earthquake, and the 1998 February floods.
When the 1906 San Francisco quake struck, little was known about earthquake engineering; the new library, then located in the main quad and nearly ready for occupancy, was destroyed, along with much of the campus (Keller, 1999). A new main library was completed in 1919, incorporating the lessons of '06 (Weber, 1999). Subsequent additions followed updated seismic research and building codes. One major addition, the West Stacks, was reinforced only months before the Loma Prieta quake.
When the campus shook on October 17, 1989, over 75,000 volumes landed on the floors of the various libraries (Weber, 1990). Nearly half the volumes in the Mathematics Library were down. Other (non-SUL) libraries on campus also suffered major damage; almost ninety-five percent of the Business Library's collection tumbled off their shelves. In general, collections on upper floors fared the worst. Bookshelves and metal shelving failed in several locations. On the other hand, some libraries had few books off their shelves. Some library equipment was damaged, including filing and microfiche cabinets and computers, but the actual physical damage to the collections turned out to be only between 2,000 and 3,000 volumes (excluding the Business Library), although book repair costs reached over $100,000. The old wing of the main library (which housed the University Archives, Special Collections, Government Documents Library, and Technical Services), was, however, badly damaged and all but the first floor and basement were deemed uninhabitable after major cracks were discovered in exterior walls and interior walls shattered.
Most staff were at work in the main library when the quake struck and took refuge under tables and in doorways. Although a few suffered minor cuts from flying glass in the older wing, almost no injuries were reported and managers were able to safely clear the area.
W. Henry (personal communication, April 25, 2001) noted that the existing emergency response plan was not used-yet it worked. The director was out of town and the associate director wasn't at all familiar with the plan and didn't know some key procedures, including the Conservation Dept.'s role in coordinating recovery, but she gathered the top managers together. She told them to manage the situation and they ably performed the task.
One of the surprises of the event was how little water damage occurred. The guidelines noted that almost all library emergencies involve water damage - and in fact a valve was knocked open in the Earth Sciences Library and showered some map cases, but no other real water damage took place within SUL itself (Weber, 1990). A hot-water pipe did break on the third floor of the Business Library and staff worked through the night in the Business Library to pull books out of the water because any contact with water would promote mold growth-one of the greatest threats to books.
Despite the need to work quickly with water damage materials, the conservation guidelines call for saving undamaged items first, if possible and performing triage, working from the best condition to the worst. Fortunately, with limited water damage, the staff was able to deal with a more or less stable situation, although safety concerns officially prevented entry into the old wing. All staff who could make it to campus the following day were asked to join teams to reshelve books. Over the next two weeks, thousands of volumes were either put back onto the shelves or sorted out for repair. Only months later after a concerted shelf reading lasting months would the collections approach their former state.
Some of the most valuable materials had been in the badly damaged old wing-Special Collections, the University Archives and administrative files-were vulnerable to rain damage through gaping walls and broken windows, but were off limits because the upper floors were ostensibly too hazardous to visit. Even so, all the collections found their way to safety within a short time by unofficial means. Henry noted that such channels often occur outside the disaster plan and help fill in the gaps (personal communication, April 25, 2001).
The lower levels of the old wing continued to be used for the next two years, although walls were sheathed in plywood for safety. Collections were shifted around or moved to remote storage while the wing was gutted and rebuilt. Although costly, the quake provided the opportunity to create a brand new library that opened to great fanfare in 1999.
Numerous floods have damaged Stanford's collections, caused by both natural events and pipe failures-or a combination of both. As noted above, most major library disasters include water. Within the last three decades at Stanford, a broken water main filled the basement of the undergraduate library, wetting over tens of thousands of volumes; the Loma Prieta quake caused water damage in two libraries; and El Niño storms have deluged collections in various libraries on two different occasions.
The most recent El Niño disaster took place on February 2-3, 1998, when heavy rainstorms flooded parts of four separate libraries. Within a short time, over 70,000 volumes were wet; by comparison, the previous major flood in the basement of the undergraduate library had claimed "only" 50,000 volumes (Buchanan, 1980). Library staff, various campus agencies and volunteers responded to middle of the night pleas and rushed to the libraries. They formed bucket brigades to move wet material out of the lower levels. Within fifteen hours, 4,000 boxes of wet materials had been shipped to a local cold-storage facility for freezing and eventual damage assessment. Once the books were removed, cleanup continued as ruined carpeting was removed, the worst of the mud and debris sucked up, damaged walls knocked out and electrical components tested. Total losses were estimated to fall between $5 and $7 million; FEMA funds would be needed to cover $1.75 million in the deductible and uninsured claims (Trei, 1998).
Had the library materials remained in the humid lower levels for more than a few hours, mold would have begun to grow; immediate carpet cleaning, mud removal and dehydration prevented mold from growing on nearby dry materials. Quick action is essential for saving damp or wet materials and once mold is established, the task becomes much more difficult. Over the next few months, conservators worked with Document Reprocessors, a firm specializing in disaster recovery, to save the damaged materials. Materials were freeze-dried, then sorted into three categories: "shelf ready," "total loss" and "books in need of repair" (Trei, 1998). In the end, only a few books were deemed a total loss and most of those were replaced.
Trei (1998) quoted W. Henry as saying he had worked from his copy of the disaster plan, but he reported that the successful recovery had much more to do with the right people having shown up at the right time (personal communication, April 25, 2001). The response had been as good as it possibly could have been, but it didn't actually depend on the plan.
With each emergency, libraries can improve the chances of preserving their own collections the next time disaster strikes. Every emergency is an enormous learning experience (Henry, personal communication, April 25, 2001). Whether in the context of a natural or human-caused disaster, familiarity with the necessary responses may mean the difference between minor and catastrophic losses. Since every emergency will occur differently than expected and procedures laid out in the best of plans will never fit the situation exactly, knowledge and flexibility are the keys to responding well.