NEHRP Site Classes for Census Tracts in Manhattan, New York City: NYCEM 2nd-Year Technical Report

NYCEM 2nd-Year Technical Report: December 31, 2000.
Earth Science Tasks: October 1, 1999 - September 30, 2000

NEHRP Site Classes for Census Tracts in Manhattan,
New York City

Lamont-Doherty Earth Observatory (LDEO) of Columbia University
P.O. Box 1000 (61 Route 9W, Seismology Building Rm 225)
Palisades NY 10964-8000
Phone: 845-365-8440
Fax: 845-365-8150
Email: jacob@ldeo.columbia.edu

List of Figures

List of Tables

Summary

The purpose of work during the second year of this planned 3-year project is to produce a census-tract-based map of the NEHRP geotechnical site classes for the entire Island of Manhattan. Geotechnical site conditions affect the shaking levels, and hence the amount of damage for a given earthquake, with softer soils generally producing higher shaking levels than stiffer site classes. The obtained map for Manhattan replaces the default site conditions which assume a relatively soft site class (Class D) for all census tracts, with what is considered a representative site condition for each census tract based on geological information which is largely based on construction-related borings. The new map is to be used by the NYCEM Consortium to test how results from the HAZUS algorithm for computing earthquake losses depend on the quality of locally derived geological and other input data for New York City.

The results indicate that most of the higher elevations in Manhattan have stiff site conditions. They belong to either the hard rock class A, or the stiff rock/soil sites of Class B that dominates the map at high elevations in Uptown Manhattan. Class B generally tends to indicate a veneer of at least 5 feet of firm soils over rock. In Midtown Manhattan the intermediate site class C is dominant. In the Lowlands of Manhattan, most but not all of them along fault zones in Upper Manhattan and in Downtown (Lower) Manhattan and at the outer fringes of the Island where man-made fills are very common, soft soil conditions of NEHRP class D are dominant. Only 2 census tracts, located on the Upper Eastside of Manhattan, have been assigned to the softest site class E. This scarcity of E sites exists despite the fact that several individual borings in other census tracts indicated E class conditions. But these conditions were not sufficiently pervasive throughout the census tract to bring the entire tract into class E.

This fact points to the nagging problem that census-tract based site class categorization is inherently problematic since the demographic constructs of census tracts do not neatly coincide with geological boundaries. We expect some additional improvements for the Manhattan site class map during the 3^rd-year effort using recently uncovered data sources, although during that final project period the focus will be on the Metropolitan region at large rather than Manhattan.

Technical Report: Site Categories for Manhattan Island, New York City

1. Introduction

This project attempts to provide census-tract-based geotechnical site classes for usage in the computer program HAZUS (NIBS, 1997) as part of the NYCEM mission to improve earthquake loss estimations for New York City. The task is to replace for each census tract the HAZUS default-value site class D with site classes A - E based on actual geologic site conditions. The work performed in the 2^nd year of this intended 3-year project uses the basic techniques developed during the 1st year (Jacob, 1999) in accordance with the NEHRP Seismic Provisions (FEMA, 1998a and b). The 1^st - Year Report and work was limited to a test area in Lower Manhattan, below 59^th Street. The 2^nd- Year Report extends this technique to the entire Island of Manhattan. In addition various improvements were included. They are:

2. Results: Map of Site Classes for Manhattan Census Tracts

The following map (Figure 1) is the main result of this 2^nd-Year project and is based on data taken from available boreholes, depth to bedrock, and existing geological surveys in Manhattan. Boring logs and depth to bedrock were obtained from MTA and NY City -DDC records (see Jacob, 1999). Geological surveys by Baskerville (1990, 1992, and 1994) and Fluhr and Murphy (1944) helped to establish underlying bedrock contours, types of rock formations, and surface topography at the grade level.

3. Boring and Other Data Locations and Density

Figure 2 shows the locations of all borehole and depth to bedrock records. Data density was greatest in lower/midtown Manhattan areas and was sparse for areas north of 137^th Street. This reflected availability of data by the MTA (Metropolitan Transportation Authority / NY Transit Authority), DDC (NY City Department of Design and Construction, see Jacob 1999) and DBR (Depth to Bedrock Records, Fluhr and Murphy, 1944). MTA records in particular were restricted by the amount of construction or renovation projects for large bus terminal and subway sites. The numbers in brackets [ ] used in the text boxes in Figure 2 represent: [1] Baskerville (1990), Sheet 3/3; [2] Baskerville, (1994), Sheet 2/2; [3] Fluhr and Murphy, (1944); [4] DDC Boreholes; [5] Depth to Bedrock; [6] MTA Boreholes.

Figures 3a-d provide more detailed views for the site-class maps and give explanations for some census tracts regarding the data sources that were used in determining the dominant site class. For census tracts that did not have one or more representative boring records, site class determinations relied on the evidence of geological surveys (Baskerville, 1990, 1992, 1994) and bedrock elevation (Fluhr and Murphy, 1944) which together with topography yields soil depth. Figure 3a demonstrates the details of this technique in census tracts above 137^thStreet. Commentary boxes explain for some of the census tracts how the assignment to the various site categories was made.

In Northern Manhattan there is a preponderance of Manhattan Schist and Inwood Marble bedrock formations. During Pleistocene time, from about 40,000 to 10,000 years ago, much of Manhattan was covered by ice. The glaciers cut away the softer Inwood marble, leaving elevated outcrops and rocky sites made primarily of Manhattan Schist. These sites are generally indicated by the firmest site class designations A and B (the latter when covered by a veneer of soils) in the schist formations extending to the Upper North West of Manhattan. They are contrasted by softer D-class sites in the Inwood Lowlands (see Figure 3a). This latter area has considerable amounts of fill and sediment; for this reason it was designated to the soft site class D.

The 125^th Street Fault Zone runs in a southeasterly strike direction from 125^thStreet near the Hudson River on the Westside of Manhattan, through the NE corner of Central Park near 110^th Street and 5^th Avenue, to East 96^th Street near the East River. It is a glacially excavated depression covered with silts, clays, swamp deposits and fills (Figure 3b). Near the southeastern end of this fault zone, just east of the NE corner of Central Park, we find the only two census tracts assigned to the softest site class E. Figure 3c shows the preponderance of soft soils of site class D in the Lowlands of Lower Manhattan. Figure 3d reproduces schematic geologic cross sections at various locations in Manhattan (Baskerville, 1994). They give the correct impression that the foundation of Manhattan is largely very firm rock. But these structural sections do not serve full justice to the fact that there are many surficial pockets of soft soils in Manhattan that can amplify the shaking during earthquakes of the buildings erected in or on these soils overlying very firm rock.

4. Interpreting MTA and DDC Geotechnical Boring Logs

Figure 4a and 4b show flow diagrams explaining the methodology for determining the site classes based on geotechnical boring logs (primarily SPT blow counts). They depict in graphic form the same scenarios (boring depth h_b in relation to total soil thickness h_s, or depth to bedrock) discussed on pages 13 and 14 of the 1^st Year NYCEM Report (Jacob, 1999). Two assumptions were changed or added for the 2^ndYear work, however. Shear wave velocity for bedrock at the bedrock/soil interface was increased as discussed below to better reflect the prevailing rock formations in the Manhattan region. The 10-ft Rule (for details, see text below) was added to ensure that soft soil layers were taken into consideration in close accordance with to NEHRP Provisions (FEMA, 1998a & b).

The decision trees of Figure 4a and 4b should be used in conjunction with Tables 1 and 2. The latter are identical to those in Jacob (1999), except for changes noted, and reflect the values assigned to various materials and site classes. Two changes were made:

(1) V_s for bedrock found in Manhattan was adjusted to higher values using the Handbook of Physical Constants (1966) as a guidance for the type of generic bedrocks typical for Manhattan; accordingly V_s for Manhattan Schist was increased to 8000 ft/sec (up from 5,000 ft/sec for bedrock used in the Year 1 report). Site classes for all DDC Borings from the first year report as well as new MTA borings were recalculated using these assumptions.

(2) The 1997 NEHRP Provisions (FEMA 1998a; Chap. 4, Sec. 4.1.2.1) do not specify SPT bore counts or shear velocity when determining whether sites qualify for the 10-ft Rule. We therefore had to redefine the Rule as applicable whenever a 10-ft layer of soils or soft clays is present with bore counts N less than ten (N<10) and shear wave velocities V_s<400 ft/s.

Table 1: Assigned velocities for materials differing from regular soils, and for which blow count information was not available.

Material	V_s (ft/s)
Any materials where drill sank under own weight (Blow count too low to measure)	100
Primary Silt	300
Top Soil (in upper 10 feet)	300 - 400
Miscellaneous Fill	400
Steel Layer with Void underneath*	500*
Brick	500
Brick and Concrete (Rubble)	500 - 700
Asphalt	1,000
Rip Rap	1,000
Concrete	3,000
Weathered Rock	3,000
Boulders	5,000
Hard Rock	8,000**

(*) eliminate if not pervasive, (**) new value chosen for Year 2 work and Report

The averages V₁₀₀ (or N₁₀₀ ) for the upper 100ft below grade were computed using Equations 6 and 7 of Jacob (1999), and the results were then categorized into site classes as follows (Table 2) :

Table 2: Definition of NEHRP Site Classes by Velocity V₁₀₀ and Blow Count N₁₀₀,

Site Class	Velocity Range	Blow Count Range
A:	V₁₀₀³ 5000 ft/s	not applicable
B:	2500 £ V₁₀₀< 5000 ft/s	N₁₀₀ ³ 100
C:	1200 £ V₁₀₀< 2500 ft/s	50 £ N₁₀₀<100
D:	600 £ V₁₀₀< 1200 ft/s	15 £ N₁₀₀<50
E:	V₁₀₀< 600 ft/s	N₁₀₀<15

5. Soil Class Determinations Using Depth to Bedrock (DBR) Only

For DBR soil borings without geotechnical data (no SPT blow counts), or for areas in which the general soil thickness was inferred from difference between topographic elevation of the grade level minus elevation of the bedrock, we use the same generic shear wave velocity function versus depth h in soils established in Jacob (1999);

V_s extends to depth h' and the remainder between h' and h = 100 ft is made of rock with shearwave velocity, V_r = 8,000 ft/s. For this combined soil-rock profile we can calculate analytically the depth-averaged shear wave velocity by dividing total distance (depth) traveled by the total travel time using the following formulae:

	V₁₀₀ = 100 ft / S t_i	(2)
where	S t_i = t_soil + t_rock	(3)
with	t_soil = ò [1/ V_s (h)] dh	(4)

for the soil depth range 0<h< h', which yields, after integration with equ. (1) for V_s (h)

whereby the term between |…..| in Equ. (5) needs to be evaluated at the upper and lower boundaries h= h' and h=0, respectively (i.e. upper minus lower value), and

Figure 6 shows the relationship between V₁₀₀ and depth to bedrock using the above-mentioned formulae. The limitations of this model were evidenced by the fact that soils could not be classified as E-type because V₁₀₀ could never be less than 600 ft/s . Site classes are associated with the following depth ranges for h', the depth of the bedrock/soil interface (rounded to the nearest 5-ft interval): Class A: 0£h'£5ft; Class B: 5<h'£15 ft; Class C: 15<h'£50 ft; Class D: 50<h'£100 ft.

6. Data-Processing of Boring Logs

The boreholes were chosen by location where information was needed and availability. Generally, borehole logs were documented following the format of Figure 6.

A database created in Microsoft Access was used to organize all data from the boring logs. Each borehole log used the cross streets of Manhattan to show the exact location of where the borehole was taken. Street corners were matched to a geocoded map (using MapInfo Professional) to find a fairly accurate latitude and longitude of each borehole.

After all of the information from the borehole logs were entered into the database, an Access report was created and analyzed in MS Excel to obtain the site class for each boring following the principles outlined in the decision trees (Figure 4a and 4b).

7. Discussion

Many of the problems outlined in the first-year report (Jacob, 1999) persisted during the extension of work to entire Manhattan Island, and some new problems were encountered. The most obvious problems were:

(2) In some census tracts the available borings with localized geotechnical information seemed unrepresentative of the overall site conditions as inferred from borings to bedrock or bedrock elevation (i.e. depth of soil column). This especially applies where the 10-ft rule required assignment of site class E, while the remainder of borings or depth to bedrock information suggested a much stiffer site class.

(3) The 10-ft rule for soft-soil site class E cannot be directly applied since definitions in the NEHRP Seismic Provisions (FEMA 1998a and b) rely on geotechnical parameters other than SPT blow counts N and shear wave velocity V_s . For this reason we had to introduce a new definition for the 10-ft rule based on V_s and N that was not included in the original NEHRP Provisions, and thus is not formally recognized by the geotechnical community. More details on this issue are discussed below.

(4) The size and shapes of census tracts impose a rather course spatial resolution and their geometries do not conform well to natural boundaries between soil classes. Hence prior production of a geologically based map of site classes would be preferable before the assignment of census tracts to site classes. This assignment could then use some weighting scheme. Such a scheme may involve weighting the site classes proportional to area, or by the exposed asset values built on respective site classes within a given census tract.

(5) We used in Upper Manhattan (above 59^th Street) the generic relation of shear velocity vs. depth of Equation (1) that had been established for Manhattan soil profiles below 59^th Street during the Year-1 work (Jacob, 1999). We did not perform any tests to see whether this relation applies to the soil profiles of Upper Manhattan, largely because of lack of calibration data in this area.

8. Conclusions

Given the geotechnical input data available at this time, we believe that the resulting census-tract-based site class map (Figure 1) is a substantial improvement over using a uniform default site class D in all of Manhattan for loss estimates with HAZUS. We believe, however, that further improvements are possible. In fact, in late 2000, while cooperating with the MTA / NY City Transit Authority (courtesy of NYCTA Chief Geologist A.N. Shah), we became aware of a new data source. It is a map that had been prepared during pre-WW-II times and is known as "Rock Data Map of Manhattan" (Scale: 1 inch in 600 feet; size about 3 by 8 feet). This map has been produced for the Borough President of Manhattan (S. Levy), by the Department of Public Works and the Civil Works Administration. The map shows the elevation of the rock floor, based on core borings and excavations at several thousand locations in Manhattan. We secured a hardcopy of this precious and rare map and have it since digitized. It is a here to fore untapped resource not included in any known digital database or scheme for determining NEHRP site categories in NY City. When its information will be included during the Year-3 phase of this project, we expect further refinements to the site class map now presented here as Figure 1. Whether or not these refinements will be producing substantially differing HAZUS loss estimates for the census-tract-based calculations for building-related losses, remains to be seen. However, once the loss estimations will be extended to lifelines and essential facilities, then such a higher-resolution, and geologically more meaningful map of site classes will probably contribute to much more accurate loss estimates for these non-building-related categories of assets.

The current results for Year 2 (largely represented by Figure 1, and the corresponding HAZUS files) have been forwarded to the Princeton University HAZUS engineering team (see Nordenson et al. 1999, for the equivalent 1st-year effort). This team carries out the HAZUS loss computations for NYCEM for Year 2 for all of Manhattan using both improved building inventories and the improved site classes reported here.

Related efforts of HAZUS-based earthquake loss estimations in the larger New York City Metropolitan region and adjacent areas, including Westchester County (McGinty and Wear, 2000) and in New Jersey (largely near Newark), were coordinated by NYCEM during this 2^nd-year effort. One major focus of this coordination was on establishing a uniform methodology for assigning NEHRP geotechnical site classes. For this purpose a "NYCEM Geotechnical Workshop" was convened at the Lamont-Doherty Earth Observatory in Palisades, NY, on March 3, 2000. It was attended by 20 practitioners of the HAZUS algorithm and by geotechnical experts from the following locations and organizations in and near the NYCEM area of interest:

Organizations: NYSEMO, NJ EMO, New York State Geological Survey, New Jersey State Geological Survey, Delaware State Geological Survey, Columbia University, CUNY, Princeton University, University of Delaware, Mueser Rutledge Geotechnical Engineers in NY City; and from the Department of Information Technology of Westchester County, NY. This workshop was preceded by extensive correspondence between the NYCEM Geotechnical Expert Group and the Geotechnical Subcommittee of the NEHRP Seismic Provisions operating under the FEMA-funded guidance of NIBS (National Institute of Building Sciences) and its BSSC (Building Seismic Safety Council). The purpose of this exchange was to fully coordinate NYCEM's interpretations of the NEHRP methodology of determining site classes with those intended by the national code committees that had created the site class definitions in the first place. During the NYCEM geotechnical workshop and the prior correspondence between NYCEM and the NEHRP Geotechnical Subcommittee it became very apparent that there will always be local conditions that will be difficult to interpret in strict terms of NEHRP site classes. This is especially so in the presence of widely differing types of input data that may not in all instances conform to NEHRP standards (i.e. may lack information on standard blow counts or shear wave velocities or even composition of soils. But this coordination effort also lead to a virtually unanimous consensus among the geotechnical experts involved. It was concluded that the remaining differences stem largely from consideration and interpretation of geotechnical input data that by their date of origin and nature do not always conform to NEHRP standards. Such auxiliary data may contribute valuable information to increase the spatial resolution. The fact that such auxiliary data may not conform to NEHRP standards will in most cases not introduce undue large uncertainties into the HAZUS loss estimates. At least they are not expected to do so in comparison to all the other sources of uncertainties inherent in the HAZUS methodology and related input data issues, for instance those concerned with the valuation and fragility of the built assets.

References

Baskerville, C.A., (1990), Bedrock and Engineering Geology Maps of New York County and Parts of Kings and Queens Counties. 3 sheets. USGS Open File Report 89-462.

Baskerville, C.A. (1992), Bedrock and Engineering Geology Maps of Bronx County and Parts of New York and Queens Counties, New York. USGS Miscellaneous Investigation Series MAPI-2003, Scale 1:24,000, 2 sheets, US Geological Survey, 1992.

Baskerville, C.A. (1994), Bedrock and Engineering Geology Maps of New York County and Parts of Kings and Queens Counties, New York, and Parts of Bergen and Hudson Counties, New Jersey. USGS Miscellaneous Investigation Series MAPI-2306, Scale 1:24,000, 2 sheets, US Geological Survey, 1994.

FEMA (1998a). 1997 Edition: NEHRP Recommended Provisions for Seismic Regulation for New Buildings, Part 1 - Provisions. Published by the Federal Emergency Management Agency (FEMA), as FEMA # 302, Washington DC

FEMA (1998b). 1997 Edition: NEHRP Recommended Provisions for Seismic Regulation for New Buildings, Part 2 - Commentary. Published by the Federal Emergency Management Agency (FEMA), as FEMA # 303, Washington DC

Fluhr, T., W. and J.J. Murphy (1944), Map showing 25-foot Contours of Bedrock of the Boro of Manhattan, City of New York, Map # 120-A. Dated 3/17/1944.

Handbook of Physical Constants (1966). Revised Edition; Sydney P. Clark Jr., Editor. The Geological Society of America. Memoir 97. Section 9: Seismic Velocities (by Frank Press) pp. 195-218;

Jacob, K.H. (1999). Site Conditions Effecting Earthquake Loss Estimates for New York City. Lamont-Doherty Earth Observatory of Columbia University, Palisades NY. May 6, 1999. 35 pages. http://www.nycem.org/techdocs/siteCondsYr1/default.asp.

NIBS (1997), Earthquake Loss Estimation Technology -HAZUS, User's Manual; prepared by the National Institute of Building Sciences (NIBS) for FEMA. Washington D.C. (1997).

McGinty, L. and S. Wear (2000). Initial Earthquake Loss Estimation Analysis for Westchester County, New York. Draft Report prepared by the Department of Information Technology, County of Westchester. http://giswww.co.westchester.ny.us

Nordenson, G., G. Deodatis, M. Tantala and A. Kampf (1999). NYCEM (New York City Area Consortium for Earthquake Loss Mitigation), 1^st Year Technical Report, May 1, 1998 to April 30, 1999: EARTHQUAKE LOSS ESTIMATION STUDY FOR THE NEW YORK CITY AREA. Department of Civil Engineering and Operations Research; PRINCETON UNIVERSITY; April 1999. http://www.nycem.org/techdocs/lossEstYr1/default.asp

Acknowledgments

We are grateful to the MTA / NYC Transit Authority to allow us generous access to their boring files. NYSEMO contributed additional funding for internship salary while MCEER administered the FEMA funding crucial for the success of this project. We would like to acknowledge especially the roles that Andrea Dargush (MCEER), Dan O'Brien (NYSEMO), Bruce Swiren (FEMA), Guy Nordenson, George Deodatis and Michael Tantala (all Princeton Univ.), and especially Dr. A. N. Shah (MTA / NYCTA) played to facilitate various aspects of this study.