(Originally published in the Fall 2011 issue of the Barron Park Association Newsletter, pp. 7-11. PDF of hardcopy in 3-column format)
In Palo Alto, much of the push for Emergency Preparedness (E-prep) has come from the neighborhood associations, especially Barron Park and Midtown. The BPA's involvement goes back at least to the early 1980s—see Inge Harding-Barlow's article in the previous BPA Newsletter (Summer 2011).
My perspective on E-prep is different from many of the other volunteers because it is a blend of several of the perspectives that you find in these groups. First, I was a victim in one of the hardest hit areas of the 1972 hurricane Agnes which produced flooding similar to that of hurricane Katrina (2005, New Orleans). Second, my professional background is Information Technology, and a substantial portion of my career involved advanced prototypes sponsored by the military. The ability to act and react faster is one of the US military's key advantages over opponents, and how it handles information is central to this capability. It is not enough to just have better information, you need to push it out faster to the people who need it, and in a form that they can readily use. When I moved on to Computer Security, I found that most of the people developing systems and procedures had no sense of the urgency and constraints imposed on someone in the middle of a cyber-attack. These experiences put me at odds with many of the other people working on E-prep in how we should approach the chaos inherent in a disaster. Third, Information Technology has shaped my approach to civics and vice versa: A large neighborhood is far more diverse in many ways than the typical online community, and accentuates various problems.
This article mentions a variety of difficulties that we have had working with the City over the years in this area. The intent is not to criticize individuals or groups, but to highlight issues of organizational culture and dynamics. I could have been much more effective if someone had given me a heads-up in this area. Understanding the implicit limits on where and how the City is effective demonstrates the critical role of volunteers, both in pushing the City and in doing tasks that the City cannot be expected to perform.
On May 17th 2004, a mountain lion had wandered deep into Palo Alto, northeast of Middlefield and Embarcadero, and was eventually shot. Although this had occurred near several elementary schools and included the time that children were walking to school, the City did not issue an alert to residents. Then, on June 6th in late morning, a bobcat was sighted first on Hubbartt Drive (below Gunn) and then in Bol Park. It was probably the bobcat that had frequently been seen in the pre-dawn hours on Roble Ridge. On this day, it may have been delayed getting home to bed, or may have been flushed from its bed. However, because of the earlier incident, this bobcat was reported as being a mountain lion. This time the City used its Teleminder telephone-based alerting system.
When I received the alert several hours later, I pushed it out to the BPA email list because I knew how slow the Teleminder system was. Back on the afternoon of Saturday July 31 1999, a walker on the bike path behind Gunn High School was attacked (weapon was a hammer), but I didn't receive the Teleminder alert until late Tuesday. Fortunately, one of the people who had witnessed the attack and intervened to drive off the attacker had then used the BPA email list to alert the neighborhood. Aside: At that time, the existence and use of a neighborhood email list had enough novelty that it made local TV news (KPIX, Channel 5).
The responses to my email indicated that there were much bigger problems with the Teleminder system
than just it being very slow:
Some residents didn't get the alert and others got garbled or incomplete messages.
I reported this to BPA E-Prep chair Patrick Muffler and he, being a scientist, had the same reaction that I did:
We need to get data on this.
He did an email survey of the neighborhood and sent the results to the City.
The problems turned out to be even bigger than we imagined.
In pursuing these problems with the City, we discovered that there were problems not just with the technology, but with the policy. First, the policy for when to use the alerting system was so highly restrictive that it was very rarely used, and thus during the heat of the situation, the police didn't remember that it was available. Second, the policy had a Catch-22: Reports by residents had to wait to be confirmed by a biologist from the state Department of Fish and Game, which would typically take several hours, but if a sighting was several hours old, it no longer qualified as urgent enough to justify an alert. As demonstrated by the episode with the misidentified bobcat, wanting confirmation was not unreasonable, but there needed to be a better balance between certainty and timeliness.
Because of constant agitation by the neighborhoods, it took only three years to get a replacement for the Teleminder system. That system, the Community Alerting and Notification System (CANS), has since been merged into AlertSCC, the county-wide version of that system. (Aside: I was responsible for the CANS acronym and associated logo). However, we failed to get the City to put in place interim measures, and that failure carries important lessons.
This effort was led by the co-chairs of Palo Alto Neighborhoods (PAN), the umbrella group for Palo Alto's neighborhood associations, plus Patrick Muffler (BPA E-prep). Those co-chairs were Annette Glanckopf Ashton of Midtown (and later Sheri Furman), Karen White of Duveneck/ St. Francis, and me. As an interim measure, we offered to the City the use of the neighborhood email lists, both for a duplicate copy of alerts being sent by Teleminder and to allow for the sending of notifications that did not qualify for Teleminder (intrusive telephone call). In rejecting our repeated offers, the City said that: (1) because not all residents used email, it was unfair to send emails to anyone, and (2) people who did use email may not want to sign up for one of the neighborhood email lists. We offered to set up a separate email list exclusively for alerts for such people, but the City rejected this. While the City's concern about equitable treatment is virtuous in most circumstances, here it resulted in paralysis. It even extended to situations that appeared to be insignificant: While the City didn't object to individual residents redistributing alerts to the various email lists, the City refused to facilitate this by putting such people at the top of the list when alerts were sent out (a feature built into the alerting software).
This process forcefully reminded me of the warped incentives—positive and negative—that government staff work under. In a normal collaboration, you want to make sure that the other parties get appropriate credit for the achievements. Yet in my dealings with a range of government agencies, not just the City, I have gotten blank stares when I raised this. Blame is another matter. A seemingly endless series of seemingly technical objections to using email made no sense to me until I realized that using email lists run by volunteers meant that the City staff would have little control but would retain responsibility, and thus would likely be doubly blamed for less-than-perfect results. On the other hand, remember that this was being proposed as an interim measure at a time the City had no viable alternative (other than hope that Teleminder didn't malfunction too badly).
There were also fundamental differences between the government culture
and
the techie culture
on how best to handle a disaster,
with the former heavily focused on top-down management and control, and
the reliability of individual components.
In contrast, one of the first things you learn as an engineer is
how to build a more reliable system out of less reliable components.
You also learn that failures are inevitable and to try to constrain them to manageable levels.
For example, my sense was that another part of the reason
the City didn't want to use multiple existing email lists for alerting was that
they saw multiple lists as requiring more management, more control and
more opportunities for something to go wrong,
and that that was worth the costs of people having to subscribe to a new list,
especially the cost
of people not getting alerts because they didn't sign up.
We techies saw the multiple lists as redundancy,
such that the failure of any one of them would be mitigated by many of its subscribers
getting the alert indirectly through people on the other lists.
As a techie and Silicon Valley denizen, I continue to be surprised by how little the concepts and attitudes underlying the Internet have penetrated the world of E-prep. After all, its core design principle was survivability and resilience (to Global Thermonuclear War), and its success is due to how easily you can add to it.
The City is haltingly moving away from its old policy of using the alerting system only in the most extreme situations and trying to find the right balance. But there have been situations that were serious enough for nearby schools to be locked down where the immediate neighbors were not alerted. For example, a resident saw the police searching her backyard and sent a query to the neighborhood email list, and found out from another resident who had gotten a text message from her daughter at a locked-down school that there was an armed robber on the loose. The alerting system should have been used to tell residents in the immediate area of this threat, to stay inside and lock their doors. Furthermore the police could have made use of the eyes and ears of those residents by telling them what to be looking for.
These incidents show one of the core dilemmas of E-prep: Emergencies are very special events and as such seem to call for special measures and equipment, but the experience has been that too often those preparations fail. Sometimes stuff works in small scale drills and tests, but doesn't scale up to full usage. Sometimes it is equipment overheating when it is in continuous use, or the equivalent for the people operating the equipment (expecting them to go without bathroom breaks or sleep). When there isn't a budget to periodically test equipment, the natural deterioration that comes with age can go unnoticed. Or it can be that the assumptions underlying the design no longer are valid. For example, the basic design of the telephone alerting systems that I have seen assumes that a household has a single telephone, and it is a landline. While these systems then cobbled on some acknowledgment of cell phones, Internet phones (VoIP), Text Messaging and email as alternatives to the landline, the design has not shifted to treating these devices as belonging to individuals who move around, rather than a household at a geographic location.
These problems are compounded by the human part of the equation.
People sometimes simply forget about the special procedures,
as in the mountain lion incident.
Other times, they simply decide not to use these special procedures
because the cost/benefit ratio is too high:
working through an unfamiliar procedure takes too much of one's mental budget
relative to what it might achieve.
When people are overwhelmed by the number of decisions that they need to make,
they try to cope by budgeting their mental resources
by not thinking as carefully as they normally would.
And mental stress compounds the problem by reducing the mental energy
you have to apply to decision-making.
Most of us refer to situations where we are over budget
as mental exhaustion,
but researchers also refer to it by terms such as
decision fatigue and ego depletion
(for those who want to pursue this topic).
In a disaster, there is inevitably going to be great stress and confusion. One important way to help people be more effective is reduce their mental deficits, thereby leaving them more energy to invest in the important decisions that they will need to make.
The first part of this is to reduce the two major sources of energy-sapping stress: uncertainty and a sense of loss of control. It is quite common to see disaster victims reduced to a near vegetative state because no one is telling them what is going on or what they can do to help. Improving Palo Alto's alerting system was an entry point into addressing this larger problem.
The second part is to reduce the number and difficulty of decisions you need to make. For example, how are you going to contact your family when the phone system is overwhelmed? And what are you going to eat, and how are you going to prepare it? If some of these predictable problems are addressed in advance, you will be able to focus more on the important things that inevitably pop up. One of easiest ways to reduce the load is to work cooperatively with the people around you. Not only can you spread out the decision-making, but the various people can contribute materials and ideas that reduce how much effort you have to dedicate to figuring out work-arounds. Conventional E-prep focuses on families in isolation from the communities around them. A major push of the current efforts is to make the community a significant part of the preparations.
The third part is to avoid the trap of special equipment and measures that won't be used. Instead you want to see where you can blend what will be important in emergency operations into your day-to-day operations as a way to keep people trained.
Although the first impression is that this is adding an unnecessary burden to normal operations, such a review can reveal problems, inefficiencies and poor priorities in those normal operations, leading to improvements.
Much of the City's communications with residents is a legacy of the days when they FAXed a press release to the newspapers, with the public reading a rehash days later. The adoption of email for distribution has been slow and erratic. The first came with a few individuals who were managing construction projects in Public Works and in Utilities: They found that certain neighborhood email lists could be a valuable supplement to the hardcopy notices they were required to put on doorsteps. This occurred primarily in Barron Park because of lobbying and prompting by me, and also because so many of the residents were on the list that notifications there had visible impacts were visibly appreciated.
The first City department to make use of the neighborhood email lists as a matter of practice was Planning, under its current Director Curtis Williams. The City Clerk has recently started making use of these lists.
One of the persistent problems is that the formatting of most of these emails
is still stuck in the days of FAXes:
logos and boilerplate at the top, and text whose layout is best suited for a sheet of paper.
I and other list managers often reformat these messages to be more suitable to how email is read,
especially on mobile devices.
The people sending these messages don't know, or remember,
that some cell phones cannot handle the PDF or MS-Word formats,
and that messages in these formats can be tens to hundreds of times larger than they need be.
This will be an important consideration when the cell networks are overloaded, as in a disaster.
Although people understand the problem, they don't see the need to change in advance:
They believe that they will remember to do such,
and that there will be no glitches.
My professional experience is very much to the contrary,
but we don't have space for my war stories
.
Part of our agenda in getting the City to send more and better messages to residents is to help change attitudes about the role of residents, especially as it would play out in a disaster. In 2004, the City's E-prep presentation was typical of most government agencies: In a disaster, we will be so overwhelmed that you should not expect to see us for days. This position even has a name: You're On Your Own (YOYO). Although a number of our City leaders were uncomfortable with this, recognizing it as an abrogation of government's basic responsibilities to its citizens, it wasn't until PAN started pushing on this that it began to change.
In conventional disaster plans, residents are viewed as little more than inanimate objects, to be gotten out of the way, and then warehoused or exported. In contrast, the common experience, including mine, is that people come together in these situations. Not only are they valuable resources, but being able to help during a disaster is important for their mental health.
Palo Alto had a forum for the various E-prep groups called the Citizen Corps Council (CCC),
but this had largely degenerated into representatives of the various groups
explaining to each other what their organizations would do in the event of a disaster.
Most were local units of national organizations that had long-established goals and practices,
and this both focused and limited what they would do.
Various of the representatives who had worked in actual disasters acknowledged that
while consistency was a virtue that helped you be more efficient and effective,
there often was an inflexibility in both their own organization and others
that crossed the line into being a foolish consistency
that impaired their effectiveness,
both in their own activities and what could be done through ad hoc cooperation.
One of the key complaints I heard from various members of the CCC was that its structure didn't facilitate getting to the level of detail that would identify these problems, and thereby encourage bridging the gaps between organizations. They were very aware of this problem because they would occasionally stumble upon instances of disconnects. For example, the hospitals and major clinics thought that the County Health Department maintained stockpiles of critical medical supplies for a disaster, and that they could easily draw upon these. Consequently, their inventories of those supplies were suitable for their normal use, and not what would be needed in a disaster. However, County Health had only a small inventory of such supplies and its disaster plan was to manage the redistribution of the (non-existent) stockpiled surpluses between hospitals. The supposition is that at one point both County Health and the hospitals had such stockpiles, but with the improved reliability of the supply chain in normal times and with budget cuts, each decided, without consulting the other, that their stockpile was redundant and could therefore be reduced. Aside: That CCC has since been disbanded and a similar organization bearing that same name has been created.
When Judy Kleinberg became mayor in 2006, her long-term interest in E-prep led her to create the Red Ribbon Task Force (RRTF), with it intended to be a results-oriented variant of the CCC. Unfortunately this was not to be. Because a mayor serves for only a year, the plan of action needs to be effectively completed before the Council break in August to allow enough time for implementation. This precluded a phased scaling up of the RRTF, resulting in too many people thrown together at one time, and it proved impossible to get past the talking phase.
The RRTF did have limited successes. First, it significantly raised the profile of E-prep. Second, it broadened the range of groups participating in E-prep, and introduced the new participants to the established organizations. For me, the critical achievement of the RRTF was that it legitimized the participation of residents in E-prep activities at the City level, with multiple neighborhood representatives being invited members of this group. The neighborhood representatives were among the most active participants of the RRTF, becoming the majority of the Executive Committee selected by Mayor Kleinberg. This was partly because we weren't constrained by a pre-defined role, but largely because our focus was on the practical questions of what residents could actually expect from these organizations during a disaster and what could be done to improve the situation.
In an unusual measure, the RRTF was extended for an additional year.
It decided to have as its focus a state-wide disaster preparedness exercise
(Golden Guardian
, November 2007).
The local portion of the drill was a bio-terror attack at a concert in San Jose
that would not be detected until after attendees had left the event,
thereby spreading the disease regionally.
The City of Palo Alto considered playing a tiny role,
practicing setting up a station to distribute medication to residents
and testing it with a few dozen volunteers.
However, it would be several more years before this could be demonstrated.
The neighborhoods took a two-pronged approach to participating. The first was the beginning of the Block Preparedness Coordinator (BPC) system, and was intentionally designed to not require any participation from the City or the other disaster response organizations, but it left open the possibility for them. However, only the California Avenue Area Development Association (CAADA) joined in this and subsequent exercises. Unfortunately, their driving force, Ronna Devincenzi, resigned as chair in 2010 in the aftermath of the City cutting down trees in the shopping district.
The second prong was to use this exercise as a vehicle to test the City's newly installed alerting system (CANS, now AlertSCC). Remembering the problems with the Teleminder system, the primary goal was to confirm not just that the calls were being placed, but that the message was actually being received and understood. Secondarily, the exercise would allow the system operators to practice with various aspects of the system in non-trivial situations, and would allow them to identify problems that actual residents might have in following the instructions in the messages. However, the person in charge of the system (now retired) was averse to testing the system, so this didn't occur.
As the RRTF was winding down, the neighborhood representatives created their own committee under the auspices of Palo Alto Neighborhoods (PAN) and chaired by Annette Glanckopf Ashton of Midtown. This committee gradually grew to include representatives of many of neighborhoods and neighborhood-like groups, such as Channing House (a neighborhood in a single large building). And other groups, such as the Red Cross and Palo Alto Medical Foundation, have participated at times.
At first, many of the members of the committee argued that the primary focus
should be on activities to get individual households better prepared for a disaster.
I argued long and hard against this, and eventually prevailed, mostly.
My argument was that such activities had been going on for decades,
and were past the point of diminishing returns,
that while there was still a need to do some of this,
the effort invested should geared to the expected small improvements.
Countering the argument
But this time it will be different!
, I couldn't
resist quoting Einstein
Insanity is doing the same thing over and over again and expecting different results.
I also pointed out that there were many established organizations already doing this,
and asked if we would have any special advantage (we didn't)
and what was to be gained by replicating those efforts.
Although this argument quickly resonated with some, it was a difficult argument for many to accept
because it required a radical shift in thinking.
The standard E-prep plan begins with
Every family must be prepared to survive on their own for 72 hours,
and anyone interested in E-prep has heard this mantra countless times.
Also many of the people involved in E-prep are there
because they are convinced it is very important, and
thus they can have a hard time understanding why
such preparations weren't a priority for everyone else.
This mantra is not only unrealistic, but has pernicious effects on disaster planning. First, it encourages the planners to make unrealistic assumptions, grossly underestimating the assistance that residents will need, and thereby encouraging them to neglect this aspect of a disaster. Second, it often inhibits measures that could reduce the very problem. A routine accompaniment of the mantra is that people who have supplies also need to be prepared to fight off their neighbors who will attempt to forcibly take those supplies for themselves (echoing the days of the family fallout shelter in the early 1960s). In real disasters, people don't isolate themselves but come together and work cooperatively, but a lot of supplies and opportunities are wasted because people are too frazzled to carefully think through plans on the spot. For example, during a power outage, an unopened refrigerator or freezer can be effective for up to several days depending on various factors (keeping food safe enough to be cooked, but not to go back to being stored). A group of families can reduce waste by opening their refrigerators one at a time, sharing among them the contents of each.
One of the problems with today's highly reliable technology is that it encourages people to become highly dependent on it while depriving them of the experience of what to do when it fails. For example, when blizzards in the eastern US knocked out electrical power for days, people were complaining about the problems of not being able to charge their cell phones. Yet they didn't think to use their cars as rechargers, both when driving around and from the battery (caution: current automotive batteries typically don't have as much reserve power as in days of yore).
I have argued that preparing such advice should be a major focus for PAN E-prep Committee because the Block Preparedness Coordinator (BPC) program would be one of the best distribution channels during a disaster. Little has been done on this because most of the effort is being put into establishing the basic BPC program, but volunteers interested in this could easily work in parallel.
The PAN E-prep committee has become too large and unwieldy to take direct action, although occasionally we forget this. Instead, it is a good venue to bring ideas to see if there is enough city-wide applicability and interest to be worth pursuing, and to recruit the necessary range of volunteers. The committee has a very diverse range of perspectives and expertise and can provides valuable critiquing and other feedback: People enthusiastic enough to pursue an idea often fail to ask the hard questions of themselves. A common situation is that people identify information that would be good to have when a disaster occurs, and focus on collecting it, but forget that it needs to be maintained so that it is reasonably up-to-date when needed. And they forget to consider whether it will be in a form that can easily be given to those that need it. These sorts of questions are a reflex for me because of professional experience. My most memorable variant of this problem was encountered in the 1980s: The US Navy would send out long-range anti-submarine patrol aircraft with highly detailed forecasts of ocean conditions, but these were often obsolete by the time the aircraft arrived in the patrol area, and with no updating capability, the crews had to proceed on their intuitions. One crew member stated that much of the paper this was printed on wound up being used only for toilet paper, and they would have greatly preferred the real thing.
In the major disaster I was in, teenage boys played an outsized role, and it was their geographically-based social network that was more important than their muscles. Within a neighborhood, they had established social and working relationships that made it easy to assemble teams that were immediately cohesive. And they knew the leaders in other neighborhoods, both near and far, facilitating larger scale cooperation and coordination. Although adults often wound up leading these larger groups, it was often the sons who were largely responsible for assembling them.
But this was a different time and place: That was almost 40 years ago—I was 21 at the time. The primary goal of the BPC program is to have social networks here that could fulfill this role in a disaster. The BPCs would be on a priority list for receiving information from the City to redistribute to residents. And they would help get information from the residents to the City and other disaster response organizations: They would know, or be told, what information needed to be collected, whom best to send it to, and how to format it so that it could readily be acted on. Most importantly, because they had already established contact with the households in their area, this would help them facilitate getting people to work together, both by encouragement and by providing suggestions on what to do. Although other leaders may well emerge on the block during a disaster, the BPC's leadership during the initial period has the important role of minimizing the social awkwardness that often stymies effective action while a group is forming. The ideal BPC would actually be a couple that was at home most of the time (so they would be there when the disaster struck). One of them would be highly social, for example, spending lots of time gardening in the front yard, and would know and be known by the neighbors. The other half of the couple would be interested in the details of E-prep.
Since the beginning of the BPC program there has been confusion with the CERT program, and questions why the BPC program should not just be part of the CERT program. Until recently, the CERTs (Community Emergency Response Teams) were called PANDAs (Palo Alto Neighborhood Disaster Activity). They are residents who have taken a series of courses that will allow them to assist the Fire Department, for example, walking an area and reporting preliminary damage assessments. During an emergency, they are to report to their assigned fire station and would be sent to where they are most needed.
In contrast, the BPCs are expected to stay on their blocks and provide coverage there. Additionally, although there is some overlap in the training for CERTs and BPCs, requiring the BPCs to have the irrelevant portions of the CERT training could be a significant impediment to recruiting appropriate BPCs.
Although the BPC program benefits from endorsement and support by the City, it is better as a grassroots program, maintaining the flexibility to adapt to what the volunteers are willing and capable of doing, rather than being constrained by a balky official decision-making process.
Historical aside:
Barron Park residents played a significant role in the creation of the PANDA program in the mid-1990s
and were a substantial portion of the first class
(Art Bayce, Katie Edwards, Inge Harding-Barlow, Jack Paddock, and Perky
Perkins).
Exercises are important for testing your plans, finding what works and what needs to be fixed. A good exercise is designed to blend both success and failure. There are portions that you are confident will work, and are intended to provide training and practice. Then there are portions that you hope will work, but need to be tested. These inevitable reveal substantial flaws and false assumptions, and importantly serve to deflate hubris. Finally, there are the experiments that you fully expect to fail, but in instructive ways that simplify creating first drafts of plans for them. Some exercises are focused on your own plans; others focus on testing the interaction with other groups.
Historical aside: In 1987, the BPA E-prep Committee worked with the City to conduct an evacuation drill in part of neighborhood (led by Inge Harding-Barlow). One of the problems highlighted was that the loudspeakers on the fire trucks were not heard by many residents, but this problem went unaddressed, occurring for real ten years later during flooding from San Francisquito Creek.
One of the problems with the typical disaster plan is that it plans to fail: It starts off with the stated expectation that the organization's normal resources and procedures will be inadequate to handle the demand created in a disaster, and then goes on to try to determine how to select and respond to the cases that it will handle, and thereby implicitly accepting that it will badly fail to meet its mission for the majority of cases. Rarely is there any attempt to think how they might quickly scale-up their capacity to better meet demand, recruiting and quickly training people and getting them resources.
For example, during a talk by a member of a medical unit sent to help after Hurricane Katrina, the speaker showed pictures of long corridors at the New Orleans basketball arena filled with people on stretchers with no attendants in sight. He said that people probably died from lack of simple attention. I asked if they had attempted to recruit locals to help with this, and pointed out that a few hundred feet away at the Superdome there were probably many such attendants, people who worked in nursing homes, hospitals and home care. His response was that his unit didn't try because their procedures had no provision for such. He mentioned that they wouldn't be able to verify the qualifications of the volunteers, another apparent example of how governmental concern about liability issues requires a level of perfection that isn't practical.
One example of scaling up would be that when traffic lights aren't functioning during a power outage to use people to direct traffic rather than have 4-way stops. My experience having done such was that an ordinary resident in good physical condition can be quite effective with only a short briefing. And having people directing traffic greatly reduces congestion, speeds emergency vehicles to their destinations, simplifies getting relief supplies in, and helps ordinary people get home. Police Chief Burns has said that he is considering how best to do this, but there are undoubtedly many more such tasks.