For a grid of data points distance "d" apart, the programs can estimate the average over two sizes of area: 1, The "16-point" estimate averages over a square of side 2d using a 4x4 array of data points (see diagram below). 2. The "9-point" estimate averages over a square of side d (shaded area below), using 4 3x3 array of data points (circled points below). Any two adjacent 16-point estimates are averages on overlapping squures, €.g., compare areas enclosed by solid and dashed lines below, Adjacent 9-point estimates are averages on non-overlapping squares which are exactly one-fourth the area of the 16-point square. The four small square averages in each large square are estimated using the same 16 points as for the large square average, 9 at a time. Data points are represented by dots on the 16-point estimate printout, and by the decimal points in the printout of 9-point estimates. The physical seales on the two printouts for a particular set of data are identical, so that the dots on the one exactly match the decimal points on the other. The solid intensity plots indicate areas above an action level by darker blocks of color (see page A-1-5). Due to the overlap on the 16-point estimates, only the smali square enclosed by the four data points in the center of an estimated square is darker when the average is above the action level. The blocks on the $-point estimates represent the true areas estimated. The intensity plots can be matched up by exactly aligning the row of asterisks (*'s) above the plot, with the first asterisk on the 16-point plot lined up on the fourtn asterisk of the 9-point plot. e @ @® © e @® ig @ ° (@) © e | iN 7, e e 1.6 Anywhere that two diagonally adjacent data points are missing, the area actually averaged over is a square with the appropriate corner knocked off (see illustration below). The purpose of this is to approximate the true shape of the island as closely as possible. FIGURE A-1-1. GRID ESTIMATES AND PLOTS A-1-3