An Analysis of Room Flooding During Inclement Weather
By Blair K. Ivey
Executive Summary
During the recent storms 18 guest rooms experienced flooding sufficient to render the rooms unsalable. I subsequently undertook a systematic analysis of the flooding in order to determine probable causes and remedies.
In the course of my investigation I discovered a high positive correlation between room flooding and the following factors: southern or western exposure, location on lower floors, attached balcony, and missing weep hole covers.
In the following sections I will examine several hypothesis related to room flooding and offer possible explanations and remedies.
Overview
A total of 54 rooms were surveyed including the 18 flooded rooms.[1] An effort was made to use a random sample but the survey sample was subject to the constraints of guest occupancy. Consequently some areas of the property are under-represented but the survey sample is large enough so that reasonable inferences can be made about the overall propensity of rooms to flood during periods of heavy rain and/or wind.
For each room surveyed the following data was obtained: exposure, with/without balcony, and the presence of weep hole covers on the window casement.
Prior to data collection and analysis four hypotheses for room flooding were considered:
1. Rooms with a southern or western exposure would be more likely to flood because the prevailing wind is out of the South and West.
2. Rooms with an attached balcony would be susceptible to flooding because the weep holes in the door casement are level with the balcony deck.
3. Lower floor rooms would be at higher risk for flooding due to water streaming down the side of the building.
4. Rooms that had weep hole covers missing from the window casement would be more likely to flood.
After discussions with Chief Engineer Steve Welton I decided not to evaluate the integrity of the caulking around room openings such as windows, PTAC, and the like because after nine years I decided that everything that could have been caulked probably had been.
Hypothesis 1:
Rooms with a southern or western exposure would be more likely to flood because the prevailing wind is out of the south and west.
Of the 54 rooms in the sample 30 faced south or west and 24 had a northern or eastern exposure. Investigation showed that 12% of the total rooms that faced south/west had water damage compared to 4% of the total rooms that faced north/east. Given that the occurrence of flooding is three times more likely in rooms that face south or west, it seems likely that the prevailing wind is a factor in room flooding.
Guest rooms with a southern or western exposure should be given priority when allocating remediation resources.
Hypothesis 2:
Rooms with an attached balcony would be susceptible to flooding because the weep holes in the door casement are level with the balcony deck.
There are 22 guest rooms with balconies on the property. Of those, 8 experienced flooding for a 36% flood rate. 47% of the 2nd floor and 14% of the 6th floor balcony rooms had some sort of flooding problem. By comparison, only 4.9% of the total non-balcony rooms flooded.
Even if the balcony drains are flowing freely water infiltration is likely in a balcony room because the door casement weep holes are flush with the balcony deck.
It may be worthwhile to plug the door casement weep holes in rooms with balconies. If it is undesirable to plug them on a permanent basis they can be caulked closed temporarily in the event of heavy rain.
Hypothesis 3:
Lower floor rooms would be at higher risk for flooding due to water streaming down the side of the building.
Of the 18 rooms that were flooded, 11 or 61% were located on the second floor, and 13 or 72% were located below the fifth floor. It seems reasonable to conclude that water flowing down the side of the building increases the risk of flooding for rooms on lower floors.
Unfortunately, there is nothing that can really be done about this problem. Other mitigation methods will have to be used.
Hypothesis 4:
Rooms that had weep hole covers missing from the window casement would be more likely to flood.
In all 44 rooms with windows only were examined including those rooms that had water damage. In no case did a room with both weep hole covers in place experience flooding, regardless of room exposure. Every room that experienced water infiltration had at least one weep hole cover missing. This indicates that missing weep hole covers is a risk factor for room flooding.
It should be noted that while none of the rooms in the sample on the 5th and 6th floors had both weep hole covers in place, only one room, Room 530, in the sample was flooded, and this room has a southern exposure. This evidence would seem to lend credence to Hypothesis 3.
Replacement of weep hole covers should be a relatively inexpensive remediation for rooms susceptible to leaking.
Conclusion
From the evidence it may be seen that guest rooms on lower floors with a southern or western exposure and equipped with a balcony and/or missing weep hole covers are at very high risk of water damage during heavy rain and wind. The pattern of water damage bears this out. The most striking example occurred in the guest rooms on the second floor with a southern exposure. Here 7 out of 9 or 78% of the rooms experienced significant water damage.
We may rank the risk factors for room flooding from highest to lowest as follows[2]:
1. Missing weep hole covers. The correlation for water damage and missing covers is 1.0. This should also be the most cost-effective remediation.
2. Room has southern or western exposure. The correlation coefficient for this circumstance is 0.78. While this problem cannot be addressed by, say, moving all of the rooms to the north or east side of the building, it does give an indication where resources should be deployed.
3. Room is located below the fifth floor. Again, the correlation coefficient for this is 0.78. And again, it is more an indication of where water damage is likely to occur rather than anything over which we have control.
4. Room has an attached balcony. The correlation here is 0.36. While this is not an especially strong correlation, it does offer value when combined with other indicators.
Most rooms at risk for water infiltration will have a combination of risk factors present. In fact, of the 54 rooms surveyed only two, 603 and 615, had only one risk factor present. They are both on the highest floor of the building, face east, do not have a balcony, and both are missing weep hole covers in the window casement.
I trust that this effort will give hotel management the information they need to evaluate and mitigate risks for guest room water damage
Appendix 1
ROOM SAMPLE
The following rooms made up the room survey sample. Rooms that experienced water damage are in bold.
202 / 408 / 603203 / 409 / 608
208 / 416 / 609
209 / 417 / 614
215 / 418 / 615
218 / 423 / 617
226 / 428 / 625
228 / 429 / 626
230 / 436 / 627
234 / 437 / 632
236 / 637
237 / 508 / 640
238 / 516
240 / 517
242 / 521
526
307 / 529
316 / 530
319 / 532
321 / 538
326
328
332
333
An effort was made to choose rooms in similar locations on each floor but room selection was driven by guest occupancy.
ROOM DISTRIBUTION BY EXPOSURE
North / East / South / West237 / 203 / 226 / 202
209 / 228 / 208
333 / 215 / 230 / 218
234
429 / 307 / 236 / 316
319 / 238
529 / 321 / 240 / 408
242 / 416
625 / 409 / 418
627 / 417 / 326
637 / 423 / 328 / 508
332 / 516
517
521 / 428 / 608
436 / 614
603
609 / 526
615 / 530
617 / 532
538
626
632
640
ROOM DISTRIBUTION BY BALCONY
Balcony / No Balcony202 / 626 / 213 / 508
203 / 640 / 228 / 516
215 / 230 / 517
218 / 234 / 521
226 / 236 / 526
240 / 237 / 529
242 / 238 / 530
532
307 / 538
316
319 / 603
321 / 608
326 / 608
328 / 614
332 / 615
333 / 617
625
408 / 627
409 / 632
416 / 637
417
418
423
428
429
436
437
Appendix 2
WEEP HOLE COVERS
In this section we can look at the data that led to the conclusions concerning weep hole covers and water infiltration risk.
This is the table for rooms in the sample that did not experience any water damage.
Two covers / One cover / No covers / Two covers / One cover / No covers208 / x / 508 / x
209 / x / 516 / x
228 / x / 517 / x
237 / x / 521 / x
526 / x
307 / x / 529 / x
316 / x / 532 / x
319 / x / 538 / x
321 / x
332 / x / 608 / x
333 / x / 609 / x
614 / x
408 / x / 617 / x
409 / x / 625 / x
416 / x / 626 / x
417 / x / 627 / x
423 / x / 632 / x
428 / .x / 637 / x
429 / x
436 / x
437 / x
Here is the data for all rooms without balconies. Again, the rooms that experienced flooding are in bold.
Two covers / One cover / No covers / Two covers / One cover / No covers208 / x / 508 / x
209 / x / 516 / x
228 / x / 517 / x
230 / x / 521 / x
234 / x / 526 / x
236 / x / 529 / x
237 / x / 530 / x
238 / x / 532 / x
538 / x
307 / x
316 / x / 603 / x
319 / x / 608 / x
321 / x / 609 / x
326 / x / 614 / x
328 / x / 615 / x
332 / x / 617 / x
333 / x / 625 / x
626 / x
408 / x / 627 / x
409 / x / 632 / x
416 / x / 637 / x
417 / x
418 / x
423 / x
428 / x
429 / x
436 / x
437 / x
Number of rooms with two covers: / 13
%: / 0.30
Number of rooms with one cover:: / 16
%: / 0.36
Number of rooms with no covers: / 15
%: / 0.34
While the probability distribution is fairly uniform, the distribution among the total rooms is not. One can see that on the 4th floor the presence of two weep hole covers per window is to be expected whereas on the 5th and 6th floors no room in the sample had both covers present. Given the overall probabilities the expectation is for each floor to look like the 2nd and 3rd floor.
Appendix 3
The correlation coefficient was determined by letting the variable under investigation (e.g. the presence of a balcony) be the independent variable and the likelihood of water infiltration be the dependent variable. The correlation coefficient was then calculated by dividing the sum of positive values for the dependent variable by the sum of positive values for the independent variable.
Ex. There are 22 rooms with balconies and 8 of them had water damage. 8 divided by 22 equals 0.36.
NOTE: The correlation coefficient for weep hole covers missing should actually be -1.0 because there is a perfect negative correlation. In other words, the lack of one or more covers does not guarantee that a room will leak (although the probability certainly goes up), but the presence of both covers virtually guarantees that the room will not leak, all other factors being equal. I decided to change the correlation coefficient to 1.0 in the interests of clarity.
[1] For a list of surveyed rooms please see Appendix I.
[2] For a discussion on how the correlation coefficients were arrived at, please see Appendix 3.