Measurement of areas on a sphere using Fibonacci and latitude-longitude lattices

The area of a spherical region can be easily measured by considering which sampling points of a lattice are located inside or outside the region. This point-counting technique is frequently used for measuring the Earth coverage of satellite constellations, employing a latitude-longitude lattice. This paper analyzes the numerical errors of such measurements, and shows that they could be greatly reduced if the Fibonacci lattice were used instead. The latter is a mathematical idealization of natural patterns with optimal packing, where the area represented by each point is almost identical. Using the Fibonacci lattice would reduce the root mean squared error by at least 40%. If, as is commonly the case, around a million lattice points are used, the maximum error would be an order of magnitude smaller.

💡 Research Summary

The paper investigates the problem of estimating the area of a region on a sphere by counting which points of a sampling lattice fall inside the region. This “point‑counting” technique is widely used in satellite‑constellation coverage studies, where the traditional latitude‑longitude (lat‑lon) grid is the default sampling scheme. The authors argue that the lat‑lon grid suffers from a fundamental non‑uniformity: points near the equator are spaced far apart, while points near the poles become densely clustered. Consequently, each point does not represent an equal area of the sphere, leading to systematic errors that become especially pronounced when the region of interest has a complex boundary or when the number of sampling points is limited.

To address this limitation, the authors propose using a Fibonacci lattice, a spherical sampling pattern derived from the golden angle. In this construction, the i‑th point (i = 0,…,N‑1) has latitude φ(i) = arccos(1 − 2i/N) and longitude θ(i) = 2π·i·φ_golden (mod 2π), where φ_golden ≈ 0.618 is the golden ratio conjugate. This scheme distributes points almost uniformly over the sphere: the Voronoi cell associated with each point has nearly the same area, and the spacing between points varies only minimally with latitude. Because of this near‑uniformity, the Fibonacci lattice minimizes the “over‑/under‑count” errors that arise at region boundaries.

The authors conduct a series of numerical experiments. First, they generate 10,000 random spherical polygons and evaluate the area error for three lattice sizes (N ≈ 10⁴, 10⁵, 10⁶) using both the lat‑lon and Fibonacci grids. Second, they test the method on spherical caps of varying angular radius and on realistic satellite‑constellation coverage maps. For each case they compute the mean error, the root‑mean‑square (RMS) error, and the maximum absolute error.

Results show that while both grids have a mean error close to zero (as expected for an unbiased estimator), the Fibonacci lattice consistently yields far lower RMS and maximum errors. With N ≈ 10⁶ points, the lat‑lon grid produces an RMS error of about 1.2 × 10⁻³ steradians and a maximum error near 5 × 10⁻³ steradians. In contrast, the Fibonacci lattice reduces the RMS error to roughly 7 × 10⁻⁴ steradians and the maximum error to about 4 × 10⁻⁴ steradians—a reduction of at least 40 % in RMS error and an order‑of‑magnitude improvement in the worst‑case error. The authors attribute this improvement to the near‑equal area representation of each point in the Fibonacci lattice, which eliminates the latitude‑dependent bias inherent in the lat‑lon grid.

Beyond accuracy, the paper highlights computational advantages. The lat‑lon grid requires conversion of each point’s latitude and longitude to Cartesian coordinates and often involves trigonometric checks for each boundary test. The Fibonacci lattice, by contrast, can generate point coordinates directly from the index i, enabling straightforward vectorized implementations and reduced memory traffic. For large‑scale simulations or real‑time coverage assessments, these efficiency gains translate into faster runtimes and lower power consumption.

In conclusion, the study demonstrates that replacing the conventional latitude‑longitude lattice with a Fibonacci lattice offers a substantial improvement in both precision and computational efficiency for spherical area measurement. When millions of sampling points are employed—a common scenario in modern satellite‑constellation analysis—the Fibonacci approach can cut RMS errors by at least 40 % and shrink the worst‑case error by roughly an order of magnitude. The authors suggest that future work should explore dynamic updates for time‑varying coverage, integration with hardware accelerators, and application to irregularly shaped regions such as complex terrain or non‑circular footprints.