GEOGRAPHICAL COORDINATE COMPUTATIONS Part I General Considerations Technical Report No. 2 ONR Task No. 389-137 Contract Nonr 1224 (48) Office of Naval Research Geography Branch W. R. Tobler Department of Geography University of Michigan Ann Arbor, Michigan December, 1964 This report has been made possible through support and sponsorship by the United States Department of the Navy, Office of Naval Research, under ONR Task Number 389-137, Contract Nonr 1224(48). Reproduction in whole or in part is permitted for any purpose by the United States Government.

REPORT AVAILABILITY NOTICE The following report has been issued by the University of Michigan under Contract Nonr-1224(48), ONR Task No. 389-137, sponsored by the Geography Branch of the Office of Naval Research. Copies are available from the Defense Documentation center for Scientific and Technical Information. GEOGRAPHICAL COORDINATE COMPUTATIONS Part I General Considerations By W. R. Tobler Technical Report Number 2 December, 1964 ABSTRACT Part I provides a discussion of the usefulness of coordinate models for studies of geographically distributed phenomena with comments on specific coordinate systems and their relevance for the analysis and inventorying of geographical information. Appendices include equations for conversion from the Public Land Survey system into latitude and longitude and to rectangular map projection coordinates. Part II considers map projections in greater detail, including estimates of the errors introduced by the substitution of map projection coordinates for spherical coordinates. Statistical computations of finite distortion are related to Tissot's Indicatrix as a general contribution to the analysis of map projections.

ACKNOWLEDGEMENTS The preparation of this report has been facilitated by the assistance of several individuals. The University of Michigan Computation Center contributed in the numerical processing, and the University's Office of Research Administration and the Geography Branch of the Office of Naval Research both provided valuable administrative advice and support. Messrs E. Franckowiak, D. Kolberg, F. Rens, and R. Yuill, graduate students in the Department of Geography, contributed in several ways and were largely responsible for the illustrations and computer programs. The project has also benefited greatly from discussions with Professors R. Berry, L. Briggs, D. Marble, and J. Nystuen.

INTRODUCT ION In recent years there has been a rapid increase in the use of formal mathematical and statistical methods for the analysis of terrestrial distributions. Such procedures have been found to be of considerable assistance in fields such as city and regional planning, demography, ecology, geography, geology, and regional science. The present study is concerned with only one of the several mathematical strategies which have been utilized for such analyses; the "coordinate model". This term is taken to include that class of studies which specifically refers to the location of observational phenomena by some system of coordinates. As an example, a technique associated with contemporary theories in geology consists of estimating the departures of empirical geological observations from a "regional trend". Here one has a collection of numerical observations (zi) at specific terrestrial locations (xi, yi), i - 1,2,..., n. The procedure begins by estimating a specific portion of the locational trend of the observations as a least squares equation z = f(x, y). The trend is then subtracted from the observations to obtain the residual, which is subsequently interpreted in terms of geological theory. In this instance the recording of locations in some system of coordinates is an essential prerequisite for the analysis. There are many other such examples. The coordinate model is widely used, and is enjoying increasing popularity since the advent of electronic computers, which permit facile manipulation of the metricized locations. The researcher now also has available to him a rapidly increasing amount of information recorded in terms of coordinates; the U.S. Bureau of the Census, for example, now provides population statistics in terms of latitude and longitude coordinates. It is difficult to over-estimate the usefulness of this manner of recording information since a large number of the analytical methods designed for the analysis of distributions assume the existence of a system of coordinates. As a system of locational labels the specific coordinates employed for the recording of observations are not of direct or inherent interest, but rather only what they enable one to deduce regarding interrelations among the phenomena observed. In this sense the particular coordinate system utilized is irrelevant. On the other hand, computations may often be simplified by the choice of a convenient coordinate notation. From a scientific point of view descriptions of phenomena and their interrelations are often simplified by appropriate formalizations involving a "natural" coordinate system for that phenomena; the use of geomagnetic coordinates in the study of terrestrial magnetism, for example. The present study, however, considers only systems which appear to be of practical utility for the large scale recording of terrestrial observations, with some emphasis on systems available in the United States. The actual surface of the earth can be referred to as the topographic surface. This bumpy two-dimensional surface is difficult to describe in all its detail. Theoretically it is possible to introduce a system of coordinates on this surface such that ground distances, etc., between all

2 points can be calculated. In practice this is not attempted. As an approximation to the topographic surface the geodesist utilizes a surface of constant gravitational potential, the geoid. This bumpy but rather smooth surface is still too complicated for practical computations. A further simplification is made by assuming the earth to have the shape of an ellipsoid, generally an ellipsoid of revolution. Geodetic coordinates are then defined for positions on this ellipsoid. An even simpler model of the topographic surface is to consider the earth to be a perfect sphere. One can continue thusly, finally arriving at the assumption that the earth is a flat plane. Each of these assumed models of the earth has its advantages and disadvantages since realism may lead to extreme cumbrousness. In practical terms, the following (somewhat contradictory) criteria seem appropriate: a) The coordinates should permit accurate and economical formulae for computation. The highest level of precision available today can be achieved only through the use of geodetic formulae. These formulae are fairly complicated. Computational simplicity can be obtained, with a consequent reduction in accuracy, by employing spherical formulae. Further computational simplicity can be achieved by the use of plane coordinates based on an appropriate map projection, again with some loss of accuracy. Computational simplicity is important for two reasons. The cost and time required for computation, particularly when large amounts of information are to be processed, can be reduced by significant amounts through the use of simplified formulae. In addition, the number of persons and agencies who can effectively make use of the information is significantly increased by the use of simplified formulae. It is more difficult to discuss accuracy, upon which the level of computational simplificty depends, since the degree of precision required in particular studies varies considerably. There is no reason to employ a method which results in accuracies greater than required or greater than those with which the information was recorded. An objective of this study has been to estimate the accuracy obtainable by employing several alternate methods. This allows the individual researcher to choose the simplest computational method which yields the requisite level of accuracy. b) A rapid and accurate method of determining the coordinates of a position should be available. In general a system of coordinates which requires a carefully executed geodetic survey can be considered highly accurate, but relatively slow. A system which enables one to read coordinates from an aerial photograph or map (either manually, mechanically, or electronically) is more rapid but the accuracy is dependent on the map scale. The convenience of this method is also dependent on the availability of maps or photographs to which the coordinate system has been affixed. Between the extremes of geodetic surveying and map scaling are a number of intermediate systems, including automatic navigation devices which permit virtually instantaneous, in situ position determination, with fair accuracy. Emphasis in this study is on map scaling procedures.

3 c) The coordinates should be widely available and should be equally convenient for use at a local, national, and international level. This objective arises since most types of information collected at a national level are used both nationally and locally. Census records provide a good example. National use of local records also is increasing. The accuracy requirements at these two levels generally differ, however. At the local level an accuracy of fifty feet may be insufficient, whereas at the national level an accuracy of five miles may suffice. TERRESTRIAL COORDINATE SYSTEMS There are many locational coordinate systems in use throughout the world. Emphasis in the current discussion is on systems available in the United States. Geodetic Coordinates: Geodetic latitude and longitude provide the traditional method of identifying locations on the surface of the earth. The earth is assumed to be an oblate spheroid and the geodetic coordinates are based on actual measurement (triangulation) between sets of locations on the topographic surface. These values are then adjusted to fit an ellipsoid representative of the region in question. Different ellipsoids are employed for the several continents of the world, with an International Ellipsoid in use for world-wide computations. Geodetic coordinates, based on the Clarke Ellipsoid of 1866 (1927 adjustment), are indicated on maps published by the U.S. Geological Survey and by the U.S. Coast and Geodetic Survey. Latitude and longitude scaled from such maps are geodetic coordinates, but such scaling will not yield the same accuracy as when the positions are established in the field by an expensive first order geodetic survey. Computations employing geodetic coordinates usually take into account the ellipsoidal shape assumed for the earth. The relevant formulae are fairly complicated. For precise geodetic work it is necessary to carry approximately fifteen significant digits. Experience with a digital computer, however, indicates that, once programmed, the ellipsoidal formulae do not require appreciably more effort than the simpler spherical formulae. A floating point program with seven significant digits (as employed) for this study) yielded values which differed less than 100 meters from more precise values over a range of 6000 kilometers. Assumptions required to apply geodetic computations to the surface of the earth are that (a) the geodetic latitudes and longitudes are known without error, (b) the ellipsoid chosen is representative of the region in question, and (c) the points involved lie on the surface of the ellipsoid (roughly, at sea level). On the other hand, this is the most accurate system available. The actual proportional error in distance, based on misclosures of the U.S. continental triangulation network, appears to be on the order of 1 D1/3 20000 where D is the computed distance in miles on the Clarke Ellipsoid of

4 1866 (1927 adjustment). The differences between the several ellipsoids in use throughout the world are small; on the order of three kilometers per 6000 miles. Connections of this length on one ellipsoidal datum are rare and the figure given does not take into account the fact that the relation between the several datums in actual use are not yet known in detail; in other words, distances between positions whose geodetic coordinates are referred to different ellipsoids may be in error by a larger figure. Astronomic Coordinates: Astronomic latitude and longitude are based on celestial observations and may depart from geodetic coordinates by as much as two kilometers at any point, due to departure of the geoid from the ellipsoid. Astronomical observations are usually available only for isolated points, and will not be considered in this report. The U.S. Public Land Survey The Public Land Survey system is based on a set of six mile squares numbered as townships north and south of a base parallel, and as ranges east and west of a base meridian. These six mile squares are then subdivided into 36 sections, each one mile square, and numbered in serpentine fashion. Each section can be further subdivided into quarter-sections, each one sixteenth of a mile in area. Several systems similar to the Public Land Survey exist in various parts of the United States; these are not considered here. Strictly speaking, the Public Land Survey is an areal identification scheme and not a metrical coordinate system, though it is often regarded as such. As a partitioning of areas the system does not differ from county or census units, except that the elemental areas are roughly of equal size and are labeled in a more convenient fashion. The system is not complete, in the sense that it is defined only for certain portions of the western United States. In these areas large amounts of information have been (and continue to be) collected and recorded in terms of Sections, Townships, and Ranges. These collections of information provide a valuable source of raw data for research workers. The Public Land Survey, however, was not designed for the analytical manipulations usually required in research work. For example, statistical analyses of spatial distributions may require calculation of the average location (and its variance, and so on) of phenomena. For such computations the distances between observed locations may be needed. The distance between the SWk, Sec. 25, T5S, R7W, Willamette Meridian, and the NEI, Sec. 2, T6N, R8E, Black Hills Meridian, is not immediately apparent, nor is there any simple formula which can be employed to obtain this difference. Observations recorded in the Public Land Survey System, however, can be convered to a coordinate system having the requisite metrical properties. This is because the Public Land Survey has many of the topological ordering properties of a coordinate system. The most direct and convenient conversion is to latitude and longitude. This can be effected in several ways. The system of Townships and Ranges is shown on U.S. Geological Survey topographic maps and approximate

5 coordinates could be scaled from there maps. A more convenient procedure is to attempt a direct calculation. The equations for such a conversion are given in the Appendix, along with an estimate of their validity. The errors are fairly small so that they might be of little consequence when working with observations from the entire United States. The urban researcher working within one city, on the other hand, might find these errors intolerable. The GEOREF and Marsden Squares Systems The GEOREF System is used by the U.S. Air Force to identify locations. It is a modification of latitude and longitude in which letters are substituted for the numerical values. Every combination of letters is taken to represent a quadrilateral bounded by latitude and longitude. In this sense the system is a partitioning of area rather than a true coordinate system. The same results can be achieved by using latitude and longitude with a convention regarding the quadrant in which the quadrilateral of area lies. The system has certain advantages in applications which require error-free rapid verbal communications (e.g. radio). The system is shown on maps published by the U.S. Air Force. The Marsden Squares system employed by the National Oceanographic Data Center is similar to the GEOREF System in that a numbering of latitude and longitude quadrilaterals is substituted for the geodetic coordinates. There are many other such systems available, including the World Aeronautical Chart designations and the International Millionth Map of the World system. The advantage of these systems is largely one of bookkeeping. Such systems are not further considered in this report. THE SPHERICAL ASSUMPTION The various computations are simplified if it is assumed that the earth is a sphere. The results of such computations do not differ by large amounts from the corresponding ellipsoidal values - the polar flattening of the earth, after all, is quite small. On the basis of a number of computations it appears that a reasonable and convenient rule of thumb is that the flattening of the earth can be taken as an approximate upper bound on the percentage error of measures calculated on a spherical as compared to an ellipsoidal assumption. This is about one part in 300. An even safer estimate is that the error will be less than one percent. For some purposes this is intolerably large, but for the majority of requirements it is far more accurate than are the data or theories now available. Detailed numerical differences between an ellipsoid and sphere for distances and angles also have recently been published. Computation of the differences for a random sample of 200 pairs of points within the continental United States resulted in the following values;

6 Distance Differences (miles) Angular differences (degrees) Average: 0.046 Average: 0.006 Standard Deviation: 1.896 Standard Deviation: 0.083 Minimum: -3.788 Minimum: -0.150 Maximum: 4.871 Maximum: 0.159 A second sample might yield somewhat different results, but the sample is probably representative for the country as a whole. As expected, the differences depend on both the distance and on the direction of the point pairs. A comparison of surface areas is given in the accompanying table. In performing these computations it has been assumed that the earth is a sphere whose radius is equal to the equatorial radius of the Clarke Ellipsoid of 1866, and that the geodetic latitude and longitude can, without modification of their numerical values, be considered to be spherical coordinates. These assumptions have the advantage of extreme simplicity. A slight improvement in accuracy can be obtained if they are not retained. For example, the spherical radius employed might be the average radius of terrestrial ellipsoid in the vicinity of the area of interest, rather than the equatorial radius. It can be proven that the average radius at any latitude is the geometric mean of the radii of curvature along the meridian and normal to the meridian, This average radius is given in the accompanying tables. Conversion of geodetic latitude to spherical latitude can also be accomplished in a large number of ways. Four of the simpler methods are illustrated in the figure. Mathematical treatments can be found in works on geodesy and map projections. THE PLANE ASSUMPTION It often is convenient to employ plane coordinates for the inventorying of analysis of terrestrially distributed phenomena. In particular, many of the numerous statistical and analytical methods which have been devised for the analysis of two dimensional distributions assume the existence of a system of Cartesian coordinates. As a very simple example, suppose that an objective is to compute the average location and the locational variance of a set of discrete phenomena on the surface of a sphere. One can proceed in several ways: a) Record the observations in latitude and longitude and then perform the calculations using the spherical formulae for average and variance. b) Plot the distribution on a map, assign arbitrary rectangular coordinates to the map, record the observations in these coordinates, and then perform the calculations using the plane formulae for the average and variance.

COMPARISON OF AREAS FOR A ONE DEGREE ZONE OF LONGITUDE WITHIN THE UNITED STATES (Values in square miles, rounded to the nearest square mile) Latitude Ellipsoidal Area Spherical Area* 26N to 27N 4265 4282 27N to 28N 4228 4244 28N to 29N 4189 4205 29N to 30N 4150 4164 30N to 31N 4109 4123 31N to 32N 4067 4080 32N to 33N 4024 4035 33N to 34N 3979 3990 34N to 35N 3934 3943 35N to 36N 3887 3895 36N to 37N 3839 3846 37N to 38N 3789 3796 38N to 39N 3739 3744 39N to 40N 3687 3692 40N to 41N 3634 3638 41N to 42N 3581 3583 42N to 43N 3526 3528 43N to 44N 3469 3471 44N to 45N 3412 3413 45N to 46N 3354 3354 46N to 47N 3295 3294 47N to 48N 3234 3232 *Radius equal to equatorial radius of Clarke ellipsoid of 1866.

CLARKE ELLIPSOID OF 1866 RADII IN MILES RADII IN KILOMETERS LATITUDE RADIUS RADIUS MEAN RADIUS RADIUS RADIUS MEAN RADIUS OF THE NORMAL RADIUS OF THE OF THE NORMAL RADIUS OF THE MERIDIAN TO THE PARALLEL MERIDIAN TO THE PARALLEL MERIDIAN MERIDIAN.0 3936.4000 3963.2258 3949.7901 3963.2258 6335.0344 6378.2064 6356.5837 6378.2064 2.5 3936.4761 3963.2513 3949.8410 3959.4791 6335.1569 6378.2474 6356.6656 6372.1767 5.0 3936.7036 3963.3276 3949.9932 3948.2459 6335.5231 6378.3703 6356.9106 6354.0986 7.5 3937.0810 3963.4543 3950.2456 3929.5463 6336.1304 6378.5741 6357.3168 6324.0045 10.0 3937.6055 3963.6303 3950.5964 3903.4138 6336.9745 6378.8574 6357.8814 6281.9481 12.5 3938.2731 3963.8542 3951.0429 3869.8950 6338.0489 6379.2178 6358.6000 6228.0048 15.0 3939.0788 3964.1245 3951.5818 3829.0502 6339.3456 6379.6528 6359.4672 6162.2714 17.5 3940.0167 3964.4391 3952.2090 3780.9528 6340.8549 6380.1591 6360.4766 6084.8658 20.0 3941.0799 3964.7957 3952.9200 3725.6892 6342.5659 6380.7330 6361.6208 5995.9276 22.5 3942.2603 3965.1915 3953.7092 3663.3592 6344.4656 6381.3699 6362.8909 5895.6169 25.0 3943.5491 3965.6236 3954.5709 3594.0755 6346.5398 6382.0652 6364.2777 5784.1154 27.5 3944.9368 3966.0886 3955.4985 3517.9636 6348.7730 6382.8137 6365.7706 5661.6249 30.0 3946.4128 3966.5832 3956.4852 3435.1618 6351.1485 6383.6097 6367.3584 5528.3680 32.5 3947.9662 3967.1036 3957.5233 3345.8212 6353.6484 6384.4472 6369.0292 5384.5881 35.0 3949.5852 3967.6458 3958.6052 3250.1051 6356.2540 6385.3198 6370.7703 5230.5477 37.5 3951.2578 3968.2058 3959.7227 3148.1893 6358.9457 6386.2209 6372.5687 5066.5297 40.0 3952.9710 3968.7792 3960.8672 3040.2613 6361.7029 6387.1439 6374.4106 4892.8360 42.5 3954.7122 3969.3619 3962.0303 2926.5205 6364.5051 6388.0815 6376.2824 4709.7877 45.0 3956.4680 3969.9492 3963.2029 2807.1780 6367.3308 6389.0268 6378.1696 4517.7242 47.5 3958.2252 3970.5369 3964.3763 2682.4558 637C.1588 6389.9725 6380.0579 4317.0029 50.0 3959.9702 3971.1203 3965.5413 2552.5869 6372.9670 6390.9114 6381.9329 4107.9987 52.5 3961.6898 3971.6950 3966.6893 2417.8147 6375.7345 6391.8364 6383.7803 3891.1034 55.0 3963.3709 3972.2567 3967.8113 2278.3928 6378.4399 6392.7403 6385.5861 3666.7252 57.5 3965.0004 3972.8010 3968.8988 2134.5844 6381.0624 6393.6163 6387.3362 3435.2876 60.0 3966.5660 3973.3239 3969.9435 1986.6619 6383.5820 6394.4577 6389.0175 3197.2289 62.5 3968.0557 3973.8212 3970.9373 1834.9064 6385.9794 6395.2581 6390.6170 2953.0015 65.0 3969.4577 3974.2891 3971.8727 1679.6072 6388.2357 6396.0112 6392.1223 2703.0712 67.5 3970.7614 3974.7242 3972.7423 1521.0611 6390.3339 6396.7114 6393.5218 2447.9155 70.0 3971.9566 3975.1230 3973.5395 1359.5722 6392.2574 6397.3531 6394.8047 2188.0237 72.5 3973.0342 3975.4824 3974.2581 1195.4507 6393.9916 6397.9316 6395.9612 1923.8952 75.0 3973.9856 3975.7997 3974.8925 1029.0127 6395.5226 6398.4422 6396.9822 1656.0388 77.5 3974.8036 3976.0724 3975.4380 860.5796 6396.8391 6398.8812 6397.8600 1384.9715 80.0 3975.4817 3976.2986 3975.8901 690.4770 6397.9304 6399.2451 6398.5876 1111.2173 82.5 3976.0147 3976.4763 3976.2454 519.0344 6398.7882 6399.5310 6399.1595 835.3065 85.0 3976.3984 3976.6042 3976.5013 346.5839 6399.4057 6399.7369 6399.5712 557.7739 87.5 3976.6297 3976.6813 3976.6555 173.4605 6399.7780 6399.8610 6399.8195 279.1581 90.0 3976.7071 3976.7071 3976.7070.0001 6399.9025 6399.9025 6399.9024.0001

INTERNATIONAL ELLIPSOID RADII IN MILES RADII IN KILOMETERS LATITUDE RADIUS RADIUS MEAN RADIUS RADIUS RADIUS MEAN RADIUS OF THE NORMAL RADIUS OF THE OF THE NORMAL RADIUS OF THE MERICIAN TO THE PARALLEL MERIDIAN TO THE PARALLEL MERIDIAN MERIDIAN.0 3936.6635 3963.3075 3949.9630 3963.3075 6335.4584 6378.3380 6356.8619 6378.3380 2.5 3936.7391 3963.3329 3950.0136 3959.5606 6335.5801 6378.3787 6356.9434 6372.3079 5.0 3936.9650 3963.4087 3950.1647 3948.3267 6335.9437 6378.5008 6357.1866 6354.2286 7.5 3937.3400 3963.5345 3950.4155 3929.6259 6336.5471 6378.7032 6357.5902 6324.1325 10.0 3937.8609 3963.7093 3950.7639 3903.4916 6337.3854 6378.9846 6358.1509 6282.0734 12.5 3938.5240 3963.9317 3951.2074 3869.9707 6338.4526 6379.3425 6358.8647 6228.1266 15.0 3939.3242 3964.2002 3951.7426 3829.1233 6339.7405 6379.7746 6359.7260 6162.3890 17.5 3940.2558 3964.5127 3952.3656 3781.0229 6341.2397 6380.2775 6360.7286 6084.9787 20.0 3941.3118 3964.8668 3953.0717 3725.7560 6342.9391 6380.8474 6361.8650 5996.0350 22.5 3942.4842 3965.2599 3953.8557 3663.4225 6344.8260 6381.4800 6363.1266 5895.7188 25.0 3943.7644 3965.6891 3954.7115 3594.1348 6346.8862 6382.1707 6364.5040 5784.2109 27.5 3945.1426 3966.1510 3955.6329 3518.0189 6349.1043 6382.9141 6365.9868 5661.7139 30.0 3946.6087 3966.6422 3956.6127 3435.2129 6351.4638 6383.7046 6367.5637 5528.4503 32.5 3948.1517 3967.1591 3957.6439 3345.8680 6353.9470 6384.5364 6369.2233 5384.6633 35.0 3949.7598 3967.6976 3958.7185 3250.1476 6356.5349 6385.4031 6370.9526 5230.6160 37.5 3951.4209 3968.2538 3959.8284 3148.2274 6359.2083 6386.2982 6372.7388 5066.5909 40.0 3953.1227 3968.8234 3960.9652 3040.2950 6361.9470 6387.2148 6374.5684 4892.8904 42.5 3954.8520 3969.4020 3962.1203 2926.5501 6364.7301 6388.1461 6376.4273 4709.8353 45.0 3956.5960 3969.9854 3963.2849 2807.2036 6367.5367 6389.0849 6378.3016 4517.7653 47.5 3958.3411 3970.5689 3964.4503 2682.4775 6370.3453 6390.0241 6380.1771 4317.0377 50.0 3960.0743 3971.1484 3965.6075 2552.6050 6373.1346 6390.9567 6382.0394 4108.0278 52.5 3961.7822 3971.7192 3966.7476 2417.8295 6375.8832 6391.8753 6383.8742 3891.1271 55.0 3963.4518 3972.2771 3967.8619 2278.4045 6378.5701 6392.7731 6385.6676 3666.7440 57.5 3965.0704 3972.8177 3968.9421 2134.5934 6381.1750 6393.6431 6387.4059 3435.3020 60.0 3966.6253 3973.3370 3969.9797 1986.6685 6383.6775 6394.4788 6389.0758 3197.2394 62.5 3968.1047 3973.8309 3970.9668 1834.9109 6386.0583 6395.2737 6390.6644 2953.0088 65.0 3969.4972 3974.2957 3971.8957 1679.6100 6388.2993 6396.0217 6392.1592 2703.0756 67.5 3970.7920 3974.7277 3972.7594 1521.0625 6390.3831 6396.7170 6393.5493 2447.9177 70.0 3971.9791 3975.1238 3973.5511 1359.5725 6392.2936 6397.3544 6394.8234 2188.0242 72.5 3973.0492 3975.4807 3974.2647 1195.4501 6394.0157 6397.9289 6395.9719 1923.8944 75.0 3973.9942 3975.7959 3974.8949 1029.0117 6395.5365 6398.4361 6396.9861 1656.0372 77.5 3974.8066 3976.0668 3975.4366 860.5784 6396.8439 6398.8721 6397.8578 1384.9695 80.0 3975.4800 3976.2913 3975.8856 690.4758 6397.9277 6399.2334 6398.5804 1111.2153 82.5 3976.0093 3976.4678 3976.2385 519.0333 6398.7796 6399.5175 6399.1484 835.3047 85.0 3976.3904 3976.5948 3976.4926 346.5831 6399.3928 6399.7219 6399.5573 557.7726 87.5 3976.6203 3976.6715 3976.6458 173.4601 6399.7629 6399.8452 6399.8040 279.1575 90.0 3976.6971 3976.6971 3976.6971.0001 6399.8864 6399.8864 6399.8864.0001

Four simple methods for the CONVERSION OF GEODETIC LATITUDE TO SPHERICAL LATITUDE _______/ /9 _ _ _ ______/ s / 9

7 c) Record the observations in latitude and longitude, apply a transformation to obtain rectangular coordinates, and then perform the claculations using the plane formulae for the average and variance. Procedure (a) has the disadvantage of being more complicated. A sufficiently small portion of the earth's surface can be considered a plane and the additional complication introduced by the use of spherical versions of the statistical formulae may not be warrented. Somewhat similar problems have been investigated in the field of land surveying and are reported in most works on geodesy. Procedures (b) and (c), above, are mathematically equivalent since maps are made by transforming latitude and longitude to plane coordinates via a map projection. Hence a study of the numerical differences between computations on a plane and on the earth becomes a study of map projection distortions. The official map producing agencies of the various countries of the world have recognized the advantages of rectangular coordinates for local purposes and save the map user the trouble of assigning his own system of rectangular coordinates. They do this by publishing maps which have the official plane coordinates printed directly on the maps. Two map projection systems of this type are available in the United States, and comparable systems exist in most other countries of the world. The use of these systems is not restricted to calculations; they might also be used to record and index information in terms of the plane coordinates, perhaps scaled from topographic maps. The systems now available have several features in common. The coordinates are usually given as rectangular coordinates, often chosen so that all values are positive. More importantly, the errors in computing as though the earth were a plane disk can be evaluated. This implies that the region within which one can perform plane computations with a specified level of precision can be defined on an a priori basis. If the allowable error is small, the region must be small or several map projection systems (called zones) must be used within the region. In the latter event conversion between zones may be required. This conversion may be directly from zone to zone or may involve reconversion to geodetic coordinates as an intermediate step. There are certain advantages in using a conformal map projection for such a system since the scale errors are then independant of direction and a scale factor can be applied to improve the accuracy of short lengths The two systems employed in the United States are: 1) The State Plane Coordinate System: This system comprises approximately 120 zones covering the entire United States, with the orientation toward individual states. The accuracy within each zone is one part in 10,000. The larger states therefore require several zones. The zones overlap,with boundaries between zones lying along minor civil divisions (usually counties). The Lambert Conformal Conical projection and the Transverse Mercator projection are employed (with only one exception),

8 depending on the shape of the individual states. This system is admirably suited to the needs of the local land surveyor and has been officially adopted by many local governmental units. In many states it has legal status, is used for land ownership, and appears on large scale maps. Conversion tables are available and simple to use for any particular zone. Conversion between zones, and especially between states, is somewhat more inconvenient. The location of the zones occasionally is awkward. In Washington state, for example, the two merging metropolitan areas of Seattle and Tacoma each lie in a separate zone. The system of State Plane Coordinates appears on all recent U.S. geological Survey topographic maps. 2) The Transverse Mercator System: Known as the Universal Transverse Mercator grid system (UTM) this system is employed by the U.S. Army. The UTM grid extends to eighty degrees north and south latitude, beyond which a Polar Stereographic grid is employed. The UTM grid extends around the world in sixty north-south zones, each covering six degrees of longitude with an overlap of one half degree. The accuracy within each zone is one part in 2500. Since different areas of the world are based on distinct ellipsoidal datums, separate tables are required for various parts of the world. Procedures are available for converting directly from onezone to adjacent zones. The zonal nature of the system is occasionally inconvenient. An alphanumeric partitioning of areas is available in the system. The UTM grid appears on all Army Map Service topographic maps, on some foreign maps, and on all recent U.S. Geological Survey topographic maps. Both of the foregoing systems have several advantages. They can be employed for virtually all computations without serious error. Further, any information recorded in either of these systems can be related to geodetic coordinates and hence to information collected anywhere else in the world. Also, these coordinates are already shown on published maps, and most photogrammetric firms are sufficiently familiar with these systems to add them to aerial photographic or maps compiled by photogrammetric methods. The disadvantages of these systems stem largely from their advantages. The very refinement required to provide coordinates of high accuracy restrict these systems to relatively small portions of the earth's surface and the transformation equations, either between zones, or to and from geodetic coordinates, are relatively complicated. These difficulties can be circumvented in several ways. When a map projection system is to be used soley for computational purposes, and not necessarily to be indicated on published maps, the choice of a particular projection depends on the type of computation contemplated. The systems cited above are so refined that they yield a stated level of accuracy for virtually all computations. This is a restriction which narrows th e of suitable projections and results in projection which require fairly involved computations. For a given

9 problem there may be a specific projection which is computationally much simpler but which yields results which are of equal accuracy. For example, a problem which requires interpolation between two points on a sphere might be attacked by using the gnomonic projection (see Appendix) since all great circles are straight lines on this projection; linear interpolation in gnomonic coordinates will yield a point lying on the arc connecting the two given points. Similarly, problems involving circles on a sphere may be attacked using the stereographic projection. In other situations computational simplicity and speed may be more important than a few tens of meters of accuracy. Kao, for example, has recently shown that the geometric (perspective) projections are especially well suited for calculation by digital computer, particularly when large amounts of locational information are required within fractions of a second (i.e., in real time problems). Clearly the choice depends on the nature of the problems and the volume of the information to be processed. Computer calclation of distance and direction on a sphere (or ellipsoid) may, in many instances, be easier than attempting to convert to plane coordinates. On the other hand a more complicated problem, as for example, occurs in weather prediction, may advantageously be solved by the use of an appropriate map projection. In this instance the problem is to construct contour-type maps of the entire northern hemisphere from information received from locations scattered within this region. Rather than attempting to solve the contour interpolation problem on a sphere, the Weather Bureau employs stereographic map projection coordinates with a local correction for the projection distortion and solves the problem in plane coordinates. If one has information recorded in latitude and longitude simple conversions to map projection coordiantes are available. For example, one can pretend that these are already the ordinate and abscissa of a plane coordinate system. The resulting projection is known as the square projection. Computations performed in this manner will differ from the true values by amounts which depend on the size of the region and on the latitude. Another simple, but slightly better, conversion is to multiply all the abscissas (longitudes) by a constant equal to the cosine of the average latitude of the region in question (the square projection with a standard parallel; also known as the rectangular projection). Such a procedure might for example, by employed in urban analysis, depending on the size of the area. Another alternative would be to transform to rectangular coordinates by converting all values into distances north and east (that is, measured along a parallel) from some arbitrary point within the region. This yields the sinusoidal projection. The equations for all of the above projections are extremely simple. Somewhat more refined, but also more complicated, solutions take into cosideration the shape of the area of concern. Albers' equal area conical projection with standard parallels at 29~ 30'N and 45~ 30'N, and Lambert's conformal conical projection with standard parallels at 33~N and 45~N, for example, are two systems which might be suitable

SQ UARE PROJECTION 1 Co - - - - -v - - - ^ ^ - - - - -^ ^ ^ ^ ^ - - - - - - F\ ^ ^?^-7! r,.

RECTANGU LAR PROJECTION [:X2CI~ my'vIiII irv Standard parallels at 60N and 60S

10 for the continental United States. The distance error in computing with these latter systems is not likely to exceed fifty miles. The use of latitude and longitude, while advantage* from the point of view of long run national needs, entails some local difficulties. In the process of recording it may be necessary to interpolate between curved lines, and the system of minutes and seconds is awkward (Decimal degrees are more convenient). The complete number of digits required to specify a given location in its world context is excessively large for local use, and the north-south and east-west designation is often superflouous (a mathematical convenience is obtained if south latitudes and west longitudes are considered negative). Finally, and perhaps most important, it is often difficult to determine the latitude and longitude of a particular spot. The direct recording and storage of geographical information in terms of rectangular coordinates circumvents some of these difficulties, but introduces others. The majority of the electro - mechanical data reduction devices (specifically, coordinate readers) which are now on the market utilize rectangular coordinates. These instruments reduce the teduim of coordinate reading, even when the desired result is latitude and longitude coordinates. In this case the inverse map projection equations are required. Curiously, these are not widely available in the literature on the subject of map projections (with a few exceptions) since the previous technology prohibited their extensive use. If the objective of the study does not include subsequent conversion to latitude and longitude, a convenient procedure is to draw arbitrary rectangular (or polar) coordinates on whatever maps or aerial photographs are available. One advantage is that this can be done by persons with no training and with virtually no intellectual effort or financial expenditure. When the map used is accurate and at a "sufficiently large" scale these arbitrary coordinates may be employed as are the map projection coordinates discussed above. If the information collected has no permanent value, this procedure is perfectly satisfactory. A disadvantage is that the errors introduced are not known. The limits within which a certain level of accuracy obtains is uncertain an one never knows whether the system can be extended to include a neighboring territory. A second major disadvantage is that it may not be possible to use information collected for ore study in a second study which either (a) encompasses a larger area than the original study, or (b) which is subsequent in time to the original study, especially if the original map has been lost, or (c) which requires a higher level of accuracy than the

11 the original study. One can imagine the difficulty of analyzing the greater metropolitan area of Kansas City if Kansas City, Missouri and Kansas City, Kansas, used two different and unrelated grid systems. Or if each bureau of a city government employed a distinct system of coordinates. The actual occurance of situations of this very nature in the field of civil engineering is what gave impetus to the establishment of the system of State Plane Coordinates by the U.S. Coast and Geodetic Survey in the 1930's. Conversion between arbitrary map coordinates can be effected with relative ease if the relation between the two systems is known, or if both systems are related to latitude and longitude by known inverse equations. If the relation between systems is not known it is theoretically possible to estimate the relation if the coordinates of a sufficient number of points are accurately known in both systems (see Appendix). Such conversions may occasionally be required but are expensive. A final distinction should be made between coordinates and areas. Coordinates describe points, not areas, and one must distinguish between an areal recording unit such as a census tract and between the coordinate system used to pinpoint some centroid taken to represent that areal unit. Areal information recording units are extremely numerous and differ widely in size and shape. As a consequence it is often necessary to convert from one areal unit (e.g. census tract) to other areal units (school district, political precint, and so on). These areal conversions differ somewhat from the coordinate conversions discussed in this report. In general, specification of the areal boundaries must be included in the mathematical conversion statements. There are then again several procedures, of varying accuracy and complexity, which may be employed for the conversions.

SELECT REFERENCES 0. S. Adams,Latitude Developments Connected with Geodesy and Cartography, Coast and Geodetic Survey Special Publication No. 67 (Washington, Government Printing Office, 1921), 132 pp. R. Bachi, "Standard Distance Measures and Related Methods for Spatial Analysis", Papers, Regional Science Assn., X(1962), pp. 83-132. G. V. Bagratuni, "On the Accuracy of Distances and Azimuths obtained from the solution of the Inverse Geodetic Problem," AERDL-T-1081, 1961. H. P. Bailey, "Two Grid Systems that Divide the Entire Surface of the Earth into Quadrilaterals of Equal Area", Transactions, American Geophysical Union, XXXVII (1956), pp. 628-635. B. Berry, "Sampling, Coding, and Storing Flood Plain Data", Agriculture Handbook No. 237, U.S. Department of Agriculture, Washington, 1962, 27 pp. B. Berry, et. al., "Geographic Ordering of Information: New Opportunities", The Professional Geographer, 16, 4 (July 1964) pp. 36-40. W. Bowie and 0. S. Adams, Grid System for Progressive Maps in the United States, U.S. Coast and Geodetic Survey Special Publication #59 (Washington, Government Printing Office, 1919), 227 pp. R. M. Brooks, Coordinate Transformation Formulas, Pacific Missile Range Technical Note. 3280-220, 1962. R. A. Bryson, "Fourier Analysis of Spatial Series," in Quantitative Geography, W. L. Garrison, ed., Forthcoming. Bureau of Land Management, Manual of Instructions for the Survey of the Public Lands of the United States (Washington, Government Printing Office, 1947), 613 pp. J. D. Carroll, Jr., Chicago Area Transportation Study, Final Report, Vol. I: Survey Findings (Chicago, CATS, 1959) 126 pp. D. Clark, Plane and Geodetic Surveying, Vol. II, 4th ed., London, Constable and Co., 1951. Coast and Geodetic Survey, Plane-Coordinate Systems, Serial 562 (Washington, Government Printing Office, 1948) 5 pp. F. H. Collins, Coordinate Transformation, Technical Report NAVTRADEVCEN 1907-7315, 1963.

R. L. Creighton, J. D. Carroll, Jr., and G. S. Finney, "Data Processing For City Planning", Journal (American Institute of Planners), XXV, 2 (1959), pp. 96-103. C. H. Deetz, and 0. S. Adams, Elements of Map Projection, Coast and Geodetic Survey Special Publication No. 68, 5th ed. (Washington, Government Printing Office, 1945), 60 pp. S. C. Dodd and F. R. Pitts, "Proposals to Develop Statistical Laws of Human Geography", Proceedings, IGU Regional Conference in Japan (Tokyo, Kasai, 1959), pp. 302-309. F. Fiala, Mathematische Kartographie, (Berlin, Verlag Technik, 1957), 316 pp. G. A. Ginzburg, "A Practical Method of Determining Distortion On Maps", Geodezist, 10 (1935), pp. 49-57. D. I. Good, "Mathematical Conversion of Section, Township, and Rxnge Notation to Cartesian Coordinates", Bulletin 170, part 3, State Geological Survey of Kansas, 1964, 30 pp. N. D. Haasbrock, Investigation of the Accuracy of Plotting and Scaling-off, Netherlands Geodetic Commission, Delft, 1955.,i T. HJgerstrand, "Statistika Primaruppgifter, Flygkartering Och'Data Processing' - Maskiner: Ett Kombineringsprojekt", Meddelanden Fran Lunds Geografiska Institution, Nr. 344 (Lund, University of Lund, 1955), pp. 233-255. E. T. Homewood, "The Computation of Geodetic Areas...", Empire Survey Review, XIII, 101, pp. 309-321. A. J. Hoskinson and J. A. Duerksen, Manual of Geodetic Astronomy, Coast and Geodetic Survey Special Publication No. 237 (Washington; Government Printing Office, 1947), 219 pp. G. L. Hosmer, Geodesy (New York, Wiley, 1946). B. R. Ingalls, Washington's Extended Use of State Plane Coordinates (Olympia, Bureau of Surveys and Maps, 1957), llpp. R. C. Kao, "Geometric Projections of the Sphere and the Spheroid", The Canadian Geographer, v. 3. (Autumn 1961), pp. 12-21. R. C. Kao, Geometric Projections and Radar Data, (Santa Monica, System Development Corp., 1959), 47 pp. R. C. Kao, "The Use of Computers in the Processing and Analysis of Geographic Information", The Geographical Review, 53(1963). pp. 530-547.

W. C. Krumbein, "Trend Surface Analysis of Contour-type Maps with Irregular Control-Point Spacing," Journal of Geophysical Research, Vol. 64, 7 (July 1959), pp. 823-834. W. D. Lambert, Effect of Variations in the Assumed Figure of the Earth on the Mapping of a Large Area, Coast and Geodetic Survey Special Publication No. 100, Serial No. 258 (Washington, Government Printing Office, 1924), 35pp. W. D. Lambert, "The Distance between two Widely Separated Points on the Surface of the Earth", Journal, Washington Academy of Sciences, XXXII,5, (1942), pp. 125-130. E. A. Lewis, "Parametric Formulas for Geodesic Curves and Distances on a Slightly Oblate Earth", Air Force Cambridge Research Laboratories, April 1963, 37 pp. A. Libault, Les Measures sur les Cartes et leur Incertitude, Paris, 1961. K. A. MacLachlan, "The Coordinate Method of 0 and D Analysis", Highway Research Board Proceedings, 29th Annual Meeting (Washington National Research Council, 1949) pp. 349-367. F. J. Marschner, "Structural Properties of Medium and small scale maps", Annals, Association of American Geographer, XXXIV, 1, pp. 1-46. F. J. Marschner, Boundaries and Records....(Washington, Farm Economics Research Division, Department of Agriculture, 1960), 73 pp. H. C. Mitchell and Lansing G. Simmons, The Plane Coordinate Systems, Coast and Geodetic Survey Special Publication No. 235, (Washington, Government Printing Office, 1945) 62 pp. F. Moser, "A Computer Oriented System in Stratigraphic Analysis", Ann Arbor, Institute of Technology, 1963. D. Neft, "Statistical Analysis for Areal Distributions", Ph.D. Thesis, Columbia University, 1962, 286 pp. S. Nordbeck, Location of Areal Data For Computer Processing, Lund Studies in Geography, Series C, 2, 1962, 41 pp. J. O'Keefe "The New Military Grid of the Department of the Army", Surveying and Mapping VIII, 4 (1948), pp. 214-216.

J. A. O'Keefe, "The Universal Transverse Mercator Grid and Projection", The Professional Geographer, N. S., IV, 5 (1952), pp. 19-24. W. D. Pattison, Beginnings of the American Rectangular Land Survey System, 1784-1800, Research paper no. 50 (Chicago; Department of Geography, University of Chicago, 1957), 248 pp. F. R. Pitts, "Committee on the Utilization of Stored Data Systems", The Professional Geographer, 16, 4 (July, 1964) pp. 41-44. Radio Technical Comission for Aeronautics, "Coordinate System Aspects of Position Identification", Journal of the Institute of Navigation, Vol. 8, #1, Spring 1961, pp. 48-58. W. F. Reynolds, Relation Between Plane Rectangular Coordinates and Geographic Positions, Coast & Geodetic Survey Special Publication #71, (Washington, G. P. P., 1936), 90 pp. E. Schmid, Transformation of Rectangular Space Coordinates, U.S. Coast and Geodetic Survey Technical Bulletin No. 15 (Washington, Government Printing Office, 1961), 13 pp. A. I. Shevanova, "On the Accuracy of Small-Scale Maps," Geodesy and Cartography, (OTS, JPRS, L-1389-D.). 1957, pp. 36-44. L. G. Simmons, "How Accurate is First-Order Traingulation?", Coast and Geodetic Survey Journal, 3 (April, 1950), pp. 53-56. B. W. Sitterly and J. A. Pierce, "Simple Computation of Distances over the Earth", Journal, Institute of Navigation, 1, 4 (Dec. 1946), pp. 62-67. E. M. Sodano, "General Non-Iterative Solution of the Inverse and Direct Geodetic Problems" paper presented at 1963 IGU meeting. Berkeley, California. State of Washington, Laws of Washington, Laws of 1945, Chapter 168, (S. B. 83), "Washington Coordinate System" (Olympia, State Printer, 1945) 2 pp. P. Thompson, Numerical Weather Analysis and Prediction, MacMillan, New York, 1961, 170 pp. W. R. Tobler, "A Comparison of Spherical and Ellipsoidal Measures", The Professional Geographer, XVI, 4 (1964), pp. 9-12. W. R. Tobler, "A Polynomial Representation of Michigan Population," 1963 Papers, Michigan Academy of Science, Arts, and Letters, XLIX (1964), pp. 445-452.

U.S. Air Force, World Geographic Reference System (GEOREF), Air Force Regulation No. 96-5, (Washington, Department of the Air Force, 1956), 7 pp. U.S. Air Force, Geodetic Distance and Azimuth Computations for Lines Under 500 Miles, ACIC Technical Report No. 59 (St. Louis, Aeronautical Chart and Information Center, 1960), 77 pp. U.S. Air Force, Geodetic Distance and Azimuth Computations for Lines Over 500 Miles, ACIC Technical Report No. 80 (St. Louis, Aeronautical Chart and Information Center, 1959), 83 pp. U.S. Air Force, Map Accuracy Evaluation, Part I, ACIC Ref. Publication No. 2, 1962. U.S. Army The Positional Accuracy of Maps, AMS Technical Report 35, 1961. U.S. Army, The Universal Grid Systems, TM 5-241 (Washington, Government Printing Office, 1951), 324 pp. U.S. Census Bureau,'National Location Code Areas", Mimeographed, 1963. C. A. Whitten, Air-Line Distances between Cities in the United States, Coast and Geodetic Survey Special Publication No. 238(Washington, Government Printing Office, 1947), 246 pp. J. R. Wray, "Photo Interpretation in Urban Area Analysis", in Manual fo Photographic Interpretation (Washington; American Society of Photogrammetry, 1960) pp. 667-716.

APPENDIX I CONVERSION FROM THE PUBLIC LAND SURVEY SYSTEM TO LATITUDE AND LONGITUDE The simplest conversion begins with a procedure which assumes that the Public Land Survey conforms to the exact specifications upon which it is based. The system, as is well known, does not conform to these specifications, for a number of reasons including measurement errors unavoidable in any empirical work and a certain laxity of supervision during the establishment of the system. For conversion into latitude and longitude the following notation is convenient: i is an index to indicate the initial point of the survey. It is necessary to distinguish at least 37 initial points in the Western United States. 0 i is the latitude of the ith base parallel. >hi is the longitude of the ith base meridian, with west longitudes negative. a is the equatorial radius of the ellipsoid taken to represent the earth. For the Clarke Ellipsoid of 1866, a=3963.2257 miles. e is the eccentricity of the ellipsoid taken to represent the earth. For the Clarke Ellipsoid of 1866 e = 0.0822718542. Mi is the radius of the meridian at the ith initial point. Mi is given by a(l - e2) Mi: - (1- e2 sin2 0p)3/2 T is the township number of the location in question, with north townships taken as positive and south townships taken as negative. R is the range number of the location in question, with east ranges taken as positive and west ranges taken as negative. Sn is the northing of the section in question, with the sign convention as above. Se is the section easting, with the sign convention as above. Qn is the quarter section northing, with signs as above. Qe is the quarter section easting. is the latitude (to be found) of the location in question.

N is the radius of curvature perpendicular to the meridian at latitude 0: N _a (1 - e2 sin2 0)1/2 is the longitude (to be found) of the location in question. The necessary equations are then: 6 T 3 + Sn + Qn 0: 0i - Mi and 6 R 3 + Se + 0Q e e N C cos The formulae are established by observing that the center of the township in question should be six miles times the number of the township north (south) of the base parallel, minus three miles to obtain the center of the township. The section northing and easting give the distance of the center of the section from the center of the township, and the quarter-section northing and easting give the distance of the center of the quarter-section from the center of the section. For the SW 1/4, Sec. 25, T 5 N, R 17 E,, one should have for example, that the center of the township is 27 miles (5 x 6 - 3) north of the initial point. Sn is -1.5 miles and Q is -0.25 miles. The total distance in the north-south direction from the initial point should therefore be + 25.25 miles. This distance must then be converted to the appropriate number of degrees and added to the latitude of the initial point. A further refinement, though hardly necessary, would be to iterate on the latitude obtained in the first step in order to adjust the meridional radius employed in the computation. Determination of the longitude is similar but slightly more difficult since the distances are measured along a parallel (a loxodrome, not a great circle or geodesic), whose radius varies with the latitude. In programming the outlined procedure for a digital computer it is simplest to employ radians instead of degrees and to store a table of Sn, Se, Qn, Qn,, 0i, and Ai. The computer can perform the assignment to the correct initial point by letter for letter examination of the name of the principal meridian. A convention is necessary to distinguish the two different initial points employed for the Fourth Principal Meridian. The method detailed assigns latitude and longitude (to about the nearest 1/4 mile) on the assumption that the Public Land Survey designations are where they should be. Of course they are not exactly there: the legal strategy is to assign to the actual locations a status of incontestable correctness, irrespective of any errors which may have been introduced

during the survey. To adjust the calculated values to conform to their legal positions requires detailed historical and empirical corrections, and can be quite tedious. For many research purposes, however, such a refinement may not be necessary. To obtain an order-of-magnitude estimate of the discrepancies, the actual latitude and longitude (as recorded on large scale topographic maps) of a scattered set of locations have been compared with the computed values. For a selection of 74 points within the State of Michigan the errors are as follows: Distribution of Errors: (N = 74) Mean: 2.849 miles Standard deviation: 1.827 miles Maximum: 9.339 miles 60o of the errors are less than 2 miles 93% of the errors are less than 5.5 miles The directional errors appear evenly distributed in all directions. A random selection of points (N = 25) from other states indicates that the errors are quite comparable and of the same order of magnitude. A sample computation is as follows: observed location: Si 1/4, Sec. 28, T 2 S, R 6 E, Michigan Meridian calculated Lat/Lon: 42~ 16'07' N, 83~0 4' 08T1 U observed Lat/Lon: 42~ 17'10' N, 63~ 44'4911" difference: 1'03' 41'i difference in miles: 2.53 direction of difference: 154.34~ (E of N) The method given above does not include an adjustment for the convergence of the meridians. Since the edges of the ranges run due north and south, the ranges become narrower as the meridians converge. To adjust for this, standard parallels are established every twenty-four miles north and south of the base parallel. The ranges are again made six miles wide at these standard parallels. The system thus is self correcting every twentyfour miles. The order of magnitude of the difference in width of ranges, separated by twenty-four miles in a north-south direction, can be established as follows: The radius of the parallel at 45~N latitude is 2807.178 miles. At a distance of 24 miles north of 45~N it is approximately 2789.834 miles. The east-west width of the northern edge of the range 24 miles north of the 45th parallel is therefore not six miles but 0.037104 miles (195.9 feet) less than six miles. On this basis the error at R 50 E, an extreme value, would be 1.86 miles. Another slight error is introduced by the topographic elevation, since the radii employed apply to a mean sea level ellipsoid.

Empirical corrections for Michigan would need to include the fact that the standard parallels are 60 (not 24) miles apart (in accord with the surveying instructions in force at the time), and that R 1 E is consistently too narrow from T 1 N to T 20 N. An adjustment for these, and other, systematic departures could be incorporated into the computer program. Conversion of the Section, Township, and Range information to latitude and longitude can be followed by conversion to map projection coordinates for map plotting or computational purposes. Direct conversion to Cartesian map coordinates also is possible but is less convenient for the entire Western United States. This is more appropriate for operations restricted to a limited area, e.g., one individual state.

1 APPENDIX II MAP PROJECTION EQUATIONS The following list gives the mathematical rules for the most commoo map projections of a sphere. The following notation is standard. Latitude of a point whose projection coordinates are desired. \ Longitude of a point whose projection coordinates are desired. ) Abscissa of a plane cartesian coordinate system. Ordinate of a plane cartesian coordinate system. C Radial distance of a plane polar coordinate system. 0 Angular direction of a plane polar coordinate system.' Latitude of the center of the map; either the point of "tangency", or a single standard parallel. lp Southerly standard parallel for projections having two standard parallels. (/ Northerly standard parallel for projections having two standard L parallels. 0 Longitude of the center of the map; either the point of "tangency", or the central meridian. C The constant of the cone for conic projections. Y The radial distance from the origin to the image of the southerly standard parallel in plane polar coordinates. All equations are given for a sphere of unit radius (R = 1) and all values are assumed to be in radians. Conversion to scale can be achieved by multiplying all distances by the appropriate scale factor. North latitudes and east longitudes are taken to be positive, i.e. i\ ~ A The equations are given in their most commonly applied form. The conical projections, for example, are not given in their oblique cases.

2 (1) Albers' equal area conic projection with two standard parallels: s9= C (X^) This puts the origin of the coordinates somewhat beyond the north pole, which is rather inconvenient. The origin can be shifted to the intersection of the southern standard parallel with the central meridian by using )= r 5sn y V I - CoSS (2) Azimuthal equidistant projection: r= arc. os I ~n' <'" +$ tosv x qoS -h: 1 A - D) J e cs\= acC S |. S \ I.f - j The origin of the coordinates is at to ) o (3) Bonne's Equal Area projection. ^- (p O-p -t, a(^ y>- 2 ) )- ) -

3 To place the origin at the intersection of the standard parallel and the central meridian use: )(z ~c^''^) \/= r S-& 5 y = r,; - lr <Lo 8 (4) Cassini Projection Ovc - - V Ic 4\I -CP t"' ~dn Co^ (X-Xo). x~~~~~ (5) Gnomonic Projection: 51iv n 5i (4\ + Cos4( Cos % a, sOC<>- X,) Stn 4 C>S 4^ - Cl x ^ ^ ^ ^s ^ - \^ 1 "~h ^~ Sin dx ) Lo-s cf CoS i0 Co 0 A- o' (6) Lambert's azimuthal equal area projection: 6 =rc o' + (26-1 Q.Co9o9 I Sini C-\ C oe= ci c ear (roec ) (7) Lambert's cylindrical equal area projection: x^ \- >^ ^ = <.~^

4 Or, with a standard parallel: X 2 ( - N) Cos TO (8) Lambert's conformal conic with two standard parallels: In Cos {, - Lh CO,,S (-z. L, ( + ) L h tA >\ ( L - ) ^Co n (s ) = lC (- 2) X = r -Y - C3 y = r, - v <os e (9) Mercator's conformal cylindrical projection: X X- X X Y: L(. if ( ) (10) Miller's Cylindrical projection: X - - Xo N-Ne

5 (11) Mollweide's equal area elliptical projection: Define chi by 1 ~/ + 3 Sin 4 1 S- sI Lf ) H, X-^.^ (A->.)(^S y X J7 s;Y(12) Orthographic projection: X= San ^ Cjs ip, - -tl< ^^ ^ eos ( \A- \. X1 Co$ s f Q (Q-(,) (13) Polyconic projection (American polyconic): r — d^~ ^ y) r +'- P C Which puts the origin at the equator. (14) Sinusoidal equal area projection: y- (>.->^) C- ^ (15) Square projection: or, with a standard parallel (also known as the rectangular projection): X — (x-y,) os cP y- CP

6 (16) Stereographic projection: 4-^ SlvCos 4- sl < (P-oiy l, $f J os Cos (>.X-.) y I s^^ S si`"o + (yt t cp Cos.p. C.s(A -A,) (17) Transverse Mercator projection. X 7 | 14 Cos Sih b^- co )( i- -o< |p ge c (.-_ <,) - ~r^

APPENDIX III LEAST SQUARES CONVERSION FROM ONE SYSTEM OF RECTANGULAR COORDINATES TO ANOTHER Given two sets of coordinates on the same map, with a minimum of five points identified in both systems of coordinates, it is possible to convert the coordinates of one set to the other by a two-dimensional version of a least-squares "line'l The procedure is most easily effected using complex numbers. Let x,y be one set of coordinates and u,v be the other set, and let W: x +- iy and Z* u — iv, where i2 -1, be the complex numbers representing the ith point. The objective is then to find the complex constants A = a,+ ia2 and B = bI+ ib2 in the equation W 1 A -I BZ such that the squared residual 2I \&.j-vw^i| J='1 is a minimum. The normal equations are readily obtained by differentiation. The equation can be rewritten as a pair of transformation equations by separating the real and imaginary parts, viz: A A Re (W) = x = a + b u + b2v A A Im (W) y a2 4- b2 u - blV /A A where x and y are the estimates of the x,y coordinates as obtained from the known u,v coordinates. The standard error, etc., of the estimate can be obtained in a manner analogous to that employed for ordinary least squares procedures.

A similar, but considerably more complicated, procedure must be employed if the two sets of coordinates do not come from the same map, or if the relation to latitude and longitude is to be estimated, or if an attempt is made to determine the map projection of an arbitrary map.

APPENDIX IV R CLARKE ELLIPSOID OF 1866 R DISTANCE AND DIRECTION / SODANO METHOD R W. R. TOBLER / UNIVERSITY OF MICHIGAN / GEOGRAPHY $COMPILE MAD, PUNCH OBJECT EXTERNAL FUNCTION (LT1.LG1*LT29LG2#DISsDIRD) R ENTRY IN RADIANS R RETURNS DISTANCE IN KILOMETERS R RETURNS DIRECTION IN DECIMAL DEGREES R ACIC TR 809 PAGES 41-47. R NECESSARY CONSTANTS VECTOR VALUES PI=314159265E-8 VECTOR VALUES TPI=628318531E-8 VECTOR VALUES ARAD=63782064E-4 VECTOR VALUES BRAD=63565838E-4 VECTOR VALUES BOVRA=9966099247E-10 VECTOR VALUES VK1=2356218428E-7 VECTOR VALUES VK2=6956258069E-5 VECTOR VALUES VK3=4986428206E-9 VECTOR VALUES VK4=-4010886986E-10 VECTOR VALUES VK5=-7994556507E-10 VECTOR VALUES VK6=3986428206E-9 VECTOR VALUES E1=17036962E-10 VECTOR VALUES E2=21769E-10 VECTOR VALUES E3=29026E-10 VECTOR VALUES E4=3628E-10 VECTOR VALUES RAD=174532925E-10 R BEGIN COMPUTATION ENTRY TO CLARKE. INDEX=1. TANB1=BOVRA*(SIN (LT1)/COS (LT1 )) TANB2=BOVRA*(SIN (LT2)/COS(LT2) ) COSB11l./SQRT. (1.+ TANB1*TANB1 ) COSB2=1./SORT.(1.+(TANB2*TANB2)) SINB =TANB 1COSB1 SINB2=TANB2*COSB2 C1=SINB1*SINB2 D1=COSB1*COSB2 DIFLON=LG2-LG1 WHENEVER DIFLON.L.O.* INDEX=-1 DIFLON=.ABS DIFLON CDIF=COS.( DIFLON) CDIS=C1+D1*CDIF SDIS=SQRT. (1-CDIS*CDIS) CA=D1*SIN (DIFLON)/SDIS CB=CA*CA CC=CDIS (1.-CB)/VK3 CD=VK4C 1 CE=VK5*C1 CF=VK6*CC CG1=2.*ATAN.(SQRT.((1.-CDIS)/(1.+CDIS))) CG=CG1/SDI S CX=CA*((CG1*(VK1+CB)+SDIS*(CC+CD)+CG*(CE+CF))/VK2) DELTAL=CX+D IFLON SDELTL=SIN. (DELTAL) CDELTL=COS. (DELTAL) DEN=TANB2*COSB1-S INB1*CDELTL

DI RATN1 ( SDELTL DEN) WHENEVER DIR.G.PIDIR=DIR-TPI DIRD=DIR/RAD DIRD=DIRD*INDEX CPHO=C1+D1*CDELTL SPHO=SQRT. ( 1.-CPHO*CPHO) CBO=D1*SDELTL/SPHO APHO=2.*ATAN (SQRT. ((1.-CPHO) / (*+CPHO))) SB02=1.-CBO*CBO C2DEL= (2 *Cl/SB02)-CPHO C4DEL= (2 *C2DEL*C2DEL) -1 SB04=SB02*SB02 S2PHO=SIN.(2.*APHO) AO=1.+E1*SB02-E2*SB04 BO=E1*SB02-E3*SB04 CO=E4*SB04 DIS=BRAD( AO*APHO+BO*SPHO*C2DEL-COS*2PHO*C4DEL) FUNCTION RETURN END OF FUNCTION

APPENDIX V R CONVERSION OF PUBLIC LAND SURVEY INFORMATION R INTO LATITUDE AND LONGITUDE R SUBROUTINES NEEDED ARE DEGRAD, RADEG, SPHEREt AVERAD R W. R. TOBLER /UNIVERSITY OF MICHIGAN / GEOGRAPHY $COMPILE MADIPRINT OBJECT.PUNCH OBJECT.EXECUTE TRC INTEGER MER.C1lC2,C3,C4,C5 INTEGER COMPAR NTWP RNGQl1Q2,SPRINCNl1 INTEGER R.ST D'N SECE(37) SECN(37),PMERID(36) BLINE(36) RMER(36) V'S DLT(1)=43.0943.0,35.0,31.0,360,961.0,64*0,34*0,40*0,40.0O 1 420.C33.0,40.0B34.0,34.0,31.0,42.0,37*0,35.0,34*0,45.0,40.0, 2 34.0,38 0,60*0,400,30.0,30.0,30.0,38O0,40*0,39*0,30eO,450O, 3 43.0 V'S MLT(1)=59*0,22.0,1.0952.0,30o.049.0,51.0,38.0,59*.O,00, 1 30.0922.0,25.0,59.0,29.0,0.0,25.0,52.0,44.0,15.0,47*.046.0, 2 7.0,28.0,7.0,0.05959.0,59.0,26.0,28.0,25.0,6.0.59.0,31.00.0 V'S SLT(1)=44.0,21*0,53.032.0,5.0,21.0*50.0,45.0922.0,50.0, 1 27.0.38.0,2.0,27,0,32.0,31.0,28.0,54.0,56.0,35.0,13*0,11.0, 2 20.0,14.0,36.0,7.0,56.0,51.0,3.0,27.0,59.0,23.0,560,11.0, 3 41.0 V'S DLG(1)=104.0,116.0,89.0,90.0,103.0,145.0,147*O091.0,84*.0 1 90.0,90.0,112.0,124.0,86.0,97.0,92.0,84.0,121.0,108.0,106.0, 2 111.0,111.0,116.0,86.0 149.0,97.0,91.088.0,84.0,89.0,109.0, 3 108.0,91.0,122.0108.0 V'S MLG(1)=3.0,23.00,14.014.0,0.0,18.0,38*.03.048.0,27*.0 1 25.0,18.0,7.0,34.0,14.024.0o21.0,54.0,31.0,53.0939.0,53.0, 2 55.0,27.0,21.0,22.09.0,1.0,16.0,8.0,56.031.09.0,944.048.0 V'S SLG(1)=16.0,35.047.0,41.0,7.0,13.026.0,7.0011.0,11.0, 1 37.0,19.0.10.0,l6.0,49.055.0,53.0,47*.059.0,12.033.0,27.0 2 17.0.21.0,24.0,8.0,36.020.038.0,540,6.0,59.0,36*.034.0, 3 49.0 SECE(0)=0.0 V'S SECE(1)=2.5i1.5,0.5,-0.5i-1.5,.-.5i-2.5i-1.5.-0.5.0.5, 11.5,2.5.2.5,1.5,0.5,-0.5-1.5-2.5,-2.5-1.5.-0.50.5.15,~ 22.5,25,1.5,05,0.5,-0.5,-1.5-2.5,-2.5i-1.5,-0.5.0.51.5.2.5 SECN(O)=0.0 V'S SECN(1)=25,25,2.255.22.5,2525.52.51.5.1.5,1.5,15,1.5.1.5 10.5,0.5.0.50.55 05,0 i50.5-0.5i-,-0.5-*-0.55-05.5 -0 5.- -1.5, 2-1.5,-1.5,-1.5-1.5,-1.5,-2.5,-2.5,-2.5-2.5,-2.5,-2.5 R1=63782064E-04 MILE=0.62136994 ESQR=6768658E-09 RAD=174532925E-10 T'H INITAL, FOR I=1,1lI.G.35 DLG(I)=-DLG(I) EXECUTE DEGRAD.(DLT(I),MLT(I),SLT(I),BLINE(I)) EXECUTE DEGRAD.(DLG(I),MLG(I),SLG(I)IPMERID(I)) SMLT=SIN.(BLINE(I)) DUM=SQRT(. 1-ESQR*SMLT*SMLT) DUMCUB=DUM P.3 DUMMY=(l -ESQR)*Rl RMER(I)=DUMMY*MILE/DUMCUB INITAL CONTINUE R'T CONS.COMPAR V'S CONS=$S3,I1*$ N=0

N1=0 P'T SKIP V'S SKIP=$1H1*$ START W'R COMPAR.GE. R'T LATLON Q1t Q2.S.T*TWP R.RNGC1 tC2,C3*C4C5, 1DLAT MLATLASLAT,DLON MLONSLON V'SLATLON=$2C1S0I 12.S3 I2.C1iS3,I2 C1.S2.5C1lS16,F3.00 12F2.0,S1,F4.02F2.0*$ EXECUTE DEGRAD.(DLATMLATSLATRLAT) EXECUTE DEGRAD.(DLON.MLONSLON.RLON) O'E R'T INDATQl,Q2,STTWP,RRNG9C1l C2,C3.C4,C5 V'SINDAT =$2C1 SS0 I2,S3 I 2C1 S3. I2C1 S2 5C1 *$ E'L N=N+1 N1=N1+1 W'R TWP.E.$N$ A=T*6.0-3.0 O'R TWP.E.$S$ A=-( T*6 0- 30)''E T'0 ERR E'L W'R RNG.E.$E$ B=R*6*0-3.0 O'R RNG.E.$W$ B=-R*6 0+3 0 O'E T'0 ERR E'L W'R Q1.E.$N$ QN=0.25 O'R Q1.E*$S$ QN=-0O25 0'R Q1.E-$ $ QN=00' E T'0 ERR E'L WIR Q2.E.$E$ QE=025 0'R Q2.E.$W$ QE=-0.25 O'R Q2.E.$ $ QE=0.O O'E T'0 ERR E'L W'R C1.E.$B$ W'R C2.E.$0$ MER=2 O'E MER=1 E'L O'R C1.E.$C$ W'R C2.E.$I$ MER=5 O'R C2.E.$0$ MER=6

O'E W'R C3.E.$I$ MER=3 O' E MER-4 E'L E'L O'R C1.E.$F$ MER=7 O'R ClES$5$ MER=8 O'R C1.E.$1$ MER=9 O'R C1.E.$4$ W'R (C4. E$A$) OR. (C5E.$A$) MER=10 O' E MER=11 E'L O'R C1.E*$G$ MER=12 O'R C1.E.$H$ W R C3E. $M$ MER=13 O0'E MER=14 E'L O'R C1.E.$I$ MER=15 0'R C1.E.$L$ MER=16 O'R C1 E.$M$ W'R C2E.$I$ MER=17 O'E MER=18 E'L O'R C1.E.$N$ W'R C2.E.$A$ MER=19 O'E MER=20 E'L O'R C1.E.$P$ MER=21 O'R C1.E.$S$ W'R C3.E.$L$ MER=22 O'R C3.E.$N$ MER=23 O'R C3.E.$W$ MER=25 O' R ( C3.E.$H$) OR.(C4.E.$H$) OR (C5.E.$H$) MER=27 0'E MER=28 E'L O'R C1.E.$2$ MER=24

O'R C1.E.$6$ MER=26 O'R C1.E.$T$ MER=29 O'R C1.E.$U$ W'R C2.E.$I$ MER=31 O' E MER=32 E'L O'R C1.E.$W$ W'R C3.E.$N$ MER=35 O'E MER=34 W'R C2.E.$A$.MER=33 E'L O'R C1.E.$3$ MER=30' E T'O ERR E'L A=A+SECN(S) +QN A=A/RMER (MER) LATIT=BLINE(MER)+A CLAT=COS.(LATIT) SMLT=SIN.(LATIT) DUM=SQRT.(1.-ESQR*SMLT*SMLT) RPAR=R1*MI LE*CLAT/DUM B=(B+SECE(S)+QE)/RPAR LONGIT=PMERID(MER)+B EXECUTE RADEG.(LATITLTD9LTMLTS) EXECUTE RADEG.(LONGIT.LGDLGMLGS) P'T ONE,NI V'S ONE=$1H ////S1.I4*$ P'TRIQ1,Q2,S,T,TWPRRNG,ClC2,C3,C4,C5, lLTD LTMLTSiLGDLGMLGS V'SRI=$1H,2C1,10H 1/4, SEC,I2,3H, TI2,C1,3H, RI2,C1. 12H,,S2,5C1,10H. MERIDIAN //S1,2(F5.0F3.0.F3OSS5), 219HCALCULATED LAT/LONG *$ W'R COMPAR.GE.1 DIFLON=LONGI T-RLON DIFLAT=LAT I T-RLAT EXECUTE RADEG.(DIFLATDLTDDLTMDLTS) EXECUTE RADEG.(DIFLONDLGDLGDDLGMDLGS) EXECUTE SPHERE.(RLATRLONLATIT,LONGITRHOALPHA) EXECUTE AVERAD.(RLATLATIT,Rl,ESQRMRAD ) RHO=RHO*MRAD ALPHA=ALPHA/RAD P'TR2 DLATMLAT,SLAT,DLOONMLONSLONDLTDDLTMDLTS DLGD 1DLGM DLGSRHOALPHA V'S R2=$1H,2(F5.0F3.0,F3.0,S5),17HOBSERVED LAT/LONG /S1,2(F 15.0,F3.0,F3.0oS5),10HDIFFERENCE /Sl,Fll.4,S6,F9.4,S6, 222HMILES AND DIRECTION *$ PUNCH FORMAT OUT,N1,RHOALPHA V'SOUT=$I5,S2,F11.4,S2,F9.4*$ E'L TRANSFER TO START ERR PRINT FORMAT ONE,N

PRINT COMMENTS THIS OBSERVATION IS INCORRECTLY RECORDED$ N1=N1-1 TRANSFER TO START E'M $COMP I LEMAD PUNCHOBJECT AVERAD EXTERNAL FUNCTION (LLTTULTR1,ESQR.R3) R MEAN RADIUS ON ELLIPSOID R LATITUDES IN RADIANS ENTRY TO AVERAD. SMLT=SIN.( (LLT+ULT)/2.) DUM=SQRT. (1.-ESQR*SMLT*SMLT) DUMC U B DUM*DUM*DUM DUMMY=(1.-ESQR )*R RMER=DUMMY/DUMCUB RPAR=R1/DUM R3=SQRT. (RMER*RPAR) FUNCTION RETURN END OF FUNCTION $COMP I LEMAD PUNCHOBJECT DEGRAD EXTERNAL FUNCTION (DEGMINSEC*RAD) R SUBROUTINE TO CONVERT DEGREES TO RADIANS ENTRY TO DEGRAD. VECTOR VALUES RADIAN=174532925E-O1 SIGN=RADIAN WHENEVER DEG.L.O.. SIGN=-RADIAN RAD=SIGN*(.ABS (DEG)+(MIN/60.)+(SEC/3600) ) FUNCTION RETURN END OF FUNCTION $COMP I LEMADPUNCHOBJECT RADEG EXTERNAL FUNCTION (RADDEGMIN*SEC) R CONVERTS RADIANS TO DEGREES,MINUTES, AND DECIMAL SECONDS INTEGER I ENTRY TO RADEG. VECTOR VALUES CONS=206264806E-3 SEC=ABS.( RAD)*CONS I=SEC/3600. REMAIN=SEC- (*3600.) DEG=I*1 I=REMAIN/60. MIN=I*1. SEC=REMAIN-( *60 ) WHENEVER RAD.L.*O*DEG=-DEG FUNCTION RETURN END OF FUNCTION $COMPILE MAD,PRINT OBJECT,PUNCH OBJECT SPHERE R COMPUTES OBLIQUE SPHERICAL COORDINATES EXTERNAL FUNCTION(NLTNLGLAT LONRHO2,GA) VECTORVALUESPI=314159265E-8 VECTORVALUESTP I =628318531E-8 VECTORVALUESPIOVR2= 157079633E-8 VECTORVALUESEPS=O.0000001 VECTORVALUESRAD= 174532925E-10

ENTRY TO SPHERE. WHENEVERNLTE. (90.*RAD) GA=LON-NLG RH02zPIOVR2-LAT OTHERWISE WHENEVER(LT.NE.NLT).OR.(LG.NE.NLG) PI=314159265E-8 TPI=2.*PI PIOVR2P I/2. EPS=0O0000001 CNLT=COS ( NLT) SNLT=SIN ( NLT) END OF CONDITIONAL WHENEVER LON.NE.LON1 DIF=LON-NLG CDIF=COS.(DIF) SDIF=SIN.( DIF) END OF CONDITIONAL CLT=COS (LAT) SLT=SIN. (LAT) Q=SLT*SNLT+CLT*CNLT*CDIF WHENEVER Q.GE.1. RH02=O0 ORWHENEVER Q.LE.-1. RH02=PI OTHERWISE RH02=ARCCO S (Q) END OF CONDITIONAL NUM=CLT*SDI F DEN=CNLT*SLT-SNLT*CLT*CDIF WHENEVER.ABS.DEN.L.EPS WHENEVER.ABS.NUM*.LEPS GA=O OTHERWISE GA=PIOVR2 WHENEVER NUM.L.O. GA=-GA END OF CONDITIONAL ORWHENEVER.ABS.NUM.L.EPS GA=O. WHENEVER DEN.L.O.,GA=PI OTHERWISE GA=ATN1.(NUMDEN) WHENEVER GA.G.PI,GAGA-TPI END OF CONDITIONAL LON 1 =LON LT=NLT LG=NLG END OF CONDITIONAL FUNCTION RETURN END OF FUNCTION NW 1/4 SEC 04 T05S R07W MICHIGAN +420405 -0850800 UNION CITY MI

DISTRIBUTION LIST (One copy unless otherwise noted) Chief of Naval Research 2 Commanding Officer Office of Naval Research Army Map Service Washington 25, D.C. 6500 Brooks Lane Attn: Geography Branch Washington 25, D.C. Defense Documentation Center 20 Dr. Reid A. Bryson Cameron Station Department of Meteorology Alexandria, Virginia 22314 University of Wisconsin Madison 6, Wisconsin Director 6 Naval Research Laboratory Mr. Robert Leland Washington 25, D.C. Cornell Aeronautical Laboratory Attn: Tech. Info. Officer P. 0. Box 235 Buffalo 21, New York Director 2 Central Intelligence Agency Dr. Richard J. Russell Washington 25, D.C. Coastal Studies Institute Attn: Map Division Louisiana State University Baton Rouge 3, Louisiana Commanding Officer Office of Naval Research Dr. Charles B. Hitchcock Branch Office American Geographical Society 230 N. Michigan Avenue Broadway at 156th Street Chicago, Illinois 60601 New York 32, New York Commanding Officer Dr. Edward B. Espenshade Office of Naval Research Department of Geography Navy No. 100 Northwestern University Fleet Post Office Evanston, Illinois New York, New York Dr. Brian J. L. Berry The Oceanographer Department of Geography U. S. Navy Oceanographic Office University of Chicago Washington 25, D.C. Chicago 37, Illinois Commanding Officer Dr. William L. Garrison U. S. Naval Photo Interpretation Department of Geography Centre Northwestern University 4301 Suitland Road Evanston, Illinois Washington 25, D.C. Dr. William C. Krumbein Geography Division Department of Geology Bureau of the Census Northwestern University Washington 25, D.C. Evanston, Illinois

DISTRIBUTION LIST (Concluded) Dr. Ruth M. Davis Professor John D. Nystuen Office of Director of Defense Department of Geography Research and Engineering The University of Michigan Department of Defense Ann Arbor, Michigan Washington 25, D.C. Professor M. F. Dacey Dr. Leslie Curry Department of Regional Science Department of Geography University of Pennsylvania Arizona State College Philadelphia 4, Pennsylvania Tempe, Arizona Professor Edwin Thomas Dr. M. Gordon Wolman Department of Geography Department of Geography Arizona State College Johns Hopkins University Tempe, Arizona Baltimore 18, Maryland Professor Forrest R. Pitts Dr. Richard C. Kao Department of Geography Economics Department University of Oregon The RAND Corporation Eugene, Oregon 1700 Main Street Santa Monica, California Professor Edwin Taaffe Department of Geography U. S. Navy Oceanographic Office The Ohio State University Washington 25, D.C. Columbus 10, Ohio Attn: Code 5005 Dr. Lewis T. Reinwald Professor J. Ross Mackay 10002 Cedar Lane Department of Geography Kensington, Maryland University of British Columbia Vancouver, British Columbia, Canada Dr. Duane F. Marble Department of Geography Professor William Bunge Northwestern University Department of Geography Evanston, Illinois Wayne State University Detroit, Michigan Dr. John C. Sherman Department of Geography Dr. Allen V. Hershey, Head University of Washington Mathematical Physics Branch Seattle 5, Washington Computation and Analysis Laboratory U. S. Naval Weapons Laboratory Dahlgren, Virginia

Unclassified Security Classification DOCUMENT CONTROL DATA - R&D (Security classeiication of title, body of abstract and indexing annotation must be entered when the overall report ie classified) 1. ORIGINATING ACTIVITY (Corporate author) 2a. REPORT SECURITY C LASSIFICATION The University of Michigan Unclassified Ann Arbor, Michigan 2b GROUP 3 REPORT TITLE GEOGRAPHICAL COORDINATE COMPUTATIONS PART I: GENERAL CONSIDERATIONS 4. DESCRIPTIVE NOTES (Type of report and inclusive datee) Technical Report No. 2 5. AUTHOR(S) (Laot name, first name, initial) Tobler, W. R. 6. REPORT DATE 7a. TOTAL NO. OF PAGES 76. NO. OF REFS December 1964 45 65 8a. CONTRACT OR GRANT NO. 9a. ORIGINATOR'S REPORT NUMBER(S) Nonr 1224(48) 05824-2-T b. PROJECT NO. c. Task No. 589-157 9b. OTHER RE PORT NO(S) (Any other number that may be assined this report) d. 10. A VA IL ABILITY/LIMITATION NOTICES Qualified requesters may obtain copies of this report from DDC. II. SUPPLEMENTARY NOTES 12. SPONSORING MILITARY ACTIVITY Office of Naval Research Geography Branch __ _ _ _ _ _ _ | ~~~~~~~Washington, D. C. 13. ABSTRACT Part I provides a discussion of the usefulness of coordinate models for studies of geographically distributed phenomena with comments on specific coordinate systems and their relevance for the analysis and inventorying of geographical information. Appendices include equations for conversion from the Public Land Survey system into latitude and longitude and to rectangular map projection coordinates. Part II considers map projections in greater detail, including estimates of the errors introduced by the substitution of map projection coordinates for spherical coordinates. Statistical computations of finite distortion are related to Tissot's Indicatrix as a general contribution to the analysis of map projections. DD, JAN 64 1473 Unclassified Security Classification

Unclassified Security Classification 14. KLINK A LINK B LINK C KEY WORDS -.ROLE WT ROLE WT ROLE WT Geography Coordinate conversion Map projections Spatial analysis Information processing INSTRUCTIONS 1. ORIGINATING ACTIVITY: Enter the name and address imposed by security classification, using standard statements of the contractor, subcontractor, grantee, Department of De- such as: fense activity or other organization (corporate author) issuing (1) Qualified requesters may obtain copies of this the report. report from DDC." 2a. REPORT SECURITY CLASSIFICATION: Enter the over- (2) "Foreign announcement and dissemination of this all security classification of the report. Indicate whether rept b D i not d "Restricted Data" is included. Marking is to be in accordance with appropriate security regulations. (3) "U. S. Government agencies may obtain copies of this report directly from DDC. Other qualified DDC 2b. GROUP: Automatic downgrading is specified in DoD Di- users shall request through rective 5200.10 and Armed Forces Industrial Manual. Enter the group number. Also, when applicable, show that optional. markings have been used for Group 3 and Group 4 as author- (4) "U. S. military agencies may obtain copies of this ized. l~~~~~~~~~~~~~ized. report directly from DDC. Other qualified users 3. REPORT TITLE: Enter the complete report title in all shall request through capital letters. Titles in all cases should be unclassified If a meaningful title cannot be selected without classification, show title classification in all capitals in parenthesis (5) "All distribution of this report is controlled. Qualimmediately following the title. ified DDC users shall request through 4. DESCRIPTIVE NOTES: If appropriate, enter the type of report, e.g., interim, progress, summary, annual, or final. If the report has been furnished to the Office of Technical Give theive e dates when a specific reporting period is Services, Department of Commerce, for sale to the public, indicovered. cate this fact and enter the price, if known. 5. AUTHOR(S): Enter the name(s) of author(s) as shown on 11. SUPPLEMENTARY NOTES: Use for additional explanaor in the report. Enter last name, first name, middle initial, tory notes. If military, show rank and branch of service. The name of the principal author is an absolute minimum requirement. 12 SPONSORING MILITARY ACTIVITY: Enter the name of the departmental project office or laboratory sponsoring (pay6. REPORT DATE: Enter the date of the report as day, ing for) the research and development. Include address. month, year; or month, year. If more than one date appears on the report, use date of publication. 13. ABSTRACT: Enter an abstract giving a brief and factual summary of the document indicative of the report, even though 7a. TOTAL NUMBER OF PAGES: The total page count it may also appear elsewhere in the body of the technical reshould follow normal pagination procedures, i.e., enter the port. If additional space is required, a continuation sheet shall number of pages containing information. b attached. 7b. NUMBER OF REFERENCES: Enter the total number of It is highly desirable that the abstract of classified reports references cited in the report. be unclassified. Each paragraph of the abstract shall end with 8a. CONTRACT OR GRANT NUMBER: If appropriate, enter an indication of the military security classification of the inthe applicable number of the contract or grant under which formation in the paragraph, represented as (TS), (S), (C), or (U) the report was written. There is no limitation on the length of the abstract. How8b, 8c, & 8d. PROJECT NUMBER: Enter the appropriate ever, the suggested length is from 150 to 225 words. military department identification, such as project number, 14. KEY WORDS: Key words are technically meaningful ter subproject number, system numbers, task number, etc..sub ject n, to. r short phrases that characterize a report and may be used as 9a. ORIGINATOR'S REPORT NUMBER(S): Enter the offi- index entries for cataloging the report. Key words must be cial report number by which the document will be identified selected so that no security classification is required. Identiand controlled by the originating activity, This number must fiers, such as equipment model designation, trade name, military be unique to this report. project code name, geographic location, may be used as key 96. OTHER REPORT NUMBE:R(S): If the report has been words but will be followed by an indication of technical conassigned any other report numbers (either by the originator The assignment of links, rules, and weights is optional or by the sponsor), also enter this number(s). 10. AVAILABILITY/LIMITATION NOTICES: Enter any limitations on further dissemination of the report, other than those Unclassified Security Classification