Case Study: Why Landsat Flies at 98 Degrees

"The overriding requirement is repeatability: the same place, the same time of day, season after season." — the design philosophy of every operational Earth-observation orbit

Executive Summary

For over fifty years, the Landsat program has photographed the entire land surface of the Earth, building the longest continuous record of our planet from space. Every one of those images owes its scientific value to a single, non-obvious number: the satellites fly at an inclination of about $98.2^\circ$ — tilted past the pole, into a slightly retrograde orbit. That number is not a choice about where to look; it is a choice about when, and it is bought entirely from the perturbation we derived this chapter, J2 oblateness. In this case study we take a real Landsat-class orbit apart and show, from the J2 nodal- regression formula alone, why $98.2^\circ$ is forced on the mission, how it delivers constant lighting for free, and what small maintenance it still needs. We are reverse-engineering a mission from its orbit — the analyst's counterpart to the designer's task in Case Study 2.

Skills applied

• Solving the J2 nodal-regression formula for the sun-synchronous inclination (§12.2).
• Distinguishing what J2 does change (the plane's orientation) from what it doesn't ($a$, $e$) (§12.2).
• Reasoning about the lighting and ground-track payoff of a precessing plane (§12.2, Ch. 9).
• Estimating the residual station-keeping and the orbit's natural stability against drag (§12.3, §12.6).
• Sanity-checking a computed orbital parameter against a real, flown mission.

Background

The vehicle and its orbit

Landsat 8 and 9 fly a near-circular, sun-synchronous orbit with these published characteristics (Tier 2 — widely reported operational values):

Why this matters

Landsat's mission is change detection: comparing an image of a place today with one from a decade ago to measure deforestation, glacier retreat, urban growth, or crop health. That comparison only works if the lighting is the same in both images — the same sun angle, the same shadows. A satellite whose orbital plane drifted relative to the Sun would photograph a place at 10 a.m. one month and 2 p.m. the next, and the shifting shadows would swamp the real changes. The entire scientific value of the archive rests on holding a constant local solar time, and that, as Chapter 9 promised and §12.2 delivered, is a job for J2.

Phase 1: What lighting requires of the orbital plane

For the sun angle to stay constant, the orbital plane must rotate to follow the Sun — one full turn per year, eastward, tracking Earth's march around the Sun:

$$ \dot\Omega_{\text{required}} = \frac{360^\circ}{365.24\ \text{days}} = 0.9856^\circ\ \text{per day, eastward} = 1.991\times10^{-7}\ \text{rad/s}. $$

This is the target. Note the direction: eastward, which (from §12.2) demands a retrograde orbit, $i > 90^\circ$. A plain polar orbit at $90^\circ$ would not precess at all and its lighting would drift through all 24 hours in a year. Landsat must lean past the pole.

Phase 2: Solving for the inclination

We now invert the J2 nodal-regression formula to find the inclination that produces exactly this drift at Landsat's altitude. With $a = 7{,}076\ \text{km}$, $e \approx 0$, the mean motion is $n = \sqrt{\mu/a^3} = 1.061\times10^{-3}\ \text{rad/s}$, and $(R_\oplus/a)^2 = (6{,}378/7{,}076)^2 = 0.812$. The formula $\dot\Omega = -\frac{3}{2}nJ_2(R_\oplus/a)^2\cos i$ rearranges to

$$ \cos i = -\frac{\dot\Omega_{\text{required}}}{\frac{3}{2}\,n J_2 (R_\oplus/a)^2} = -\frac{1.991\times10^{-7}}{\frac{3}{2}(1.061\times10^{-3})(1.0826\times10^{-3})(0.812)} = -\frac{1.991\times10^{-7}}{1.399\times10^{-6}} = -0.1423. $$

$$ i = \arccos(-0.1423) = \boxed{98.2^\circ}. $$

That is Landsat's inclination, derived from nothing but the demand for constant lighting and Earth's oblateness. The mission did not pick $98.2^\circ$ and discover it was sun-synchronous; it demanded sun- synchronicity and the J2 formula handed it $98.2^\circ$. Let us confirm the round trip in code.

import math
MU, RE, J2 = 3.986e5, 6378.0, 1.0826e-3   # km^3/s^2, km, -

def sun_sync_inclination(a):
    n = math.sqrt(MU / a**3)
    raan_req = math.radians(360.0 / 365.2422) / 86400.0    # rad/s, eastward
    return math.degrees(math.acos(-raan_req / (1.5 * n * J2 * (RE/a)**2)))

print(round(sun_sync_inclination(7076.0), 2))   # Landsat altitude 705 km
# Expected output:
# 98.18

The $98.18^\circ$ we compute rounds to the $98.2^\circ$ Landsat actually flies — a sanity check that the physics of §12.2 is not just plausible but quantitatively the reason the satellite is tilted the way it is.

🔧 Engineering Reality: the inclination tracks the altitude. If a follow-on Landsat were placed a little higher, its sun-synchronous inclination would be a little larger (a higher orbit feels a weaker bulge, so it must lean more to get the same $0.9856^\circ$/day). This is why every sun-synchronous mission has its own specific inclination between about $96^\circ$ and $102^\circ$, keyed to its exact altitude — there is no single "SSO inclination," only a curve of them that the formula above traces.

Phase 3: The payoff — constant lighting and a repeating ground track

With the plane precessing at $0.9856^\circ$/day, the angle between the orbit and the Sun stays fixed, so Landsat crosses the equator (heading south) at about $10{:}00$ a.m. local time on every pass, year after year. Consider the counterfactual: without J2 precession, the plane would hold fixed in inertial space while the Earth–Sun line swept around it, and the equator-crossing time would drift by the full $24\ \text{hours}$ over one year — about $2\ \text{hours}$ per month. Within a single season the 10 a.m. scenes would become afternoon scenes, and the archive would be worthless for change detection. J2 buys the entire mission concept.

A second gift falls out of the altitude. Landsat's period is $T = 2\pi\sqrt{a^3/\mu} = 98.7\ \text{minutes}$, giving $1{,}440/98.7 = 14.6$ orbits per day. Over $16\ \text{days}$ that is $233\ \text{orbits}$ — a whole number — so after 16 days the ground track exactly repeats and every point on Earth has been imaged under the same 10 a.m. light. The altitude was tuned to make the repeat cycle close on an integer, and the inclination was tuned (via J2) to fix the lighting. Two perturbation-aware choices, one orbit, a fifty-year data record.

🔗 Connection: J2 changes the plane, not the altitude. Notice what we did not have to correct. J2's secular effect is entirely on $\Omega$ (and $\omega$); it does not change $a$ or $e$, so Landsat's altitude and near-circular shape are stable under oblateness. That is exactly the §12.2 result — "J2 rotates the orbit's orientation, it does not resize it" — and it is why a sun-synchronous orbit, once established, holds its lighting essentially for free.

Phase 4: What still needs maintaining

Sun-synchronicity is free, but Landsat is not maintenance-free. Two small residual jobs remain:

• Frozen-orbit and ground-track upkeep. Higher gravity harmonics and drag slowly perturb the tiny eccentricity and the exact ground-track position. Landsat performs occasional small maneuvers (a few m/s per year) to hold its "frozen" eccentricity and keep the 16-day repeat pattern locked. These are station-keeping burns in the sense of §12.6, just a modest budget.
• Drag make-up — but barely. At $705\ \text{km}$, atmospheric drag is far weaker than at the ISS's $400\ \text{km}$ (density falls steeply with altitude, §12.3), so the natural orbital lifetime is decades to a century, not months. Landsat needs only tiny, infrequent altitude trims rather than the constant reboosting the ISS demands. Its altitude was chosen partly so that drag is a nuisance, not a threat.

🐛 Find the Error. An analyst inspecting a Landsat two-line element set sees the RAAN change by about $15^\circ$ over 15 days and concludes "the orbit is decaying — the plane is falling toward the equator." What has the analyst confused?

Answer

Nothing is decaying. A RAAN change of $\sim 15^\circ$ over $15\ \text{days}$ is exactly $\sim 1^\circ$ per day — the designed sun-synchronous nodal regression ($0.9856^\circ$/day), the intended eastward precession of the plane that keeps the lighting constant. It is a rotation of the plane's orientation ($\Omega$), not a change in altitude or a "fall." Decay would show up as a shrinking semi-major axis (a rising mean motion), not a drifting node. The analyst mistook the mission's core design feature for a failure — a reminder that in a perturbed orbit, a changing "constant" element is often working exactly as intended.

Phase 5: Sanity check and synthesis

Every number closes. The inclination we derived from J2 ($98.18^\circ$) matches the flown value ($98.2^\circ$); the period we computed ($98.7\ \text{min}$) gives the integer $233$-orbit, 16-day repeat cycle Landsat advertises; and the altitude explains both the long drag lifetime and the fine resolution that makes Landsat imagery useful. The mission is a perturbation, harnessed: Earth's imperfect shape, which a naive designer would treat as an error to fight, is instead the free motor that turns Landsat's orbital plane once a year and makes fifty years of comparable imagery possible. This is the sun-synchronous payoff Chapter 9 promised, now audited on a real spacecraft.

Discussion Questions

1. Landsat's inclination ($98.2^\circ$) is derived from a requirement on lighting, yet it also happens to give near-global coverage. Explain how one inclination serves both "when" and "where," and why they are not the same requirement (recall §9.5).
2. If a Landsat successor flew at $600\ \text{km}$ instead of $705\ \text{km}$, would its sun-synchronous inclination be larger or smaller? Reason from the altitude dependence before computing.
3. J2 holds Landsat's lighting for free but does nothing for its eccentricity or exact ground track. Why do those still need active maintenance, and roughly what delta-v scale does that imply?
4. A dawn–dusk sun-synchronous orbit (§9.5) keeps the satellite in near-perpetual sunlight. What inclination and local-time choice produces it, and what subsystem (Chapter 25) benefits most?

Your Turn: Extensions

• Option A (analysis). Compute the sun-synchronous inclination for altitudes of $400$, $600$, $800$, and $1{,}000\ \text{km}$, and plot inclination versus altitude. Confirm the curve rises with altitude and explain the trend physically.
• Option B (computation). Write a function local_time_drift(a, i) that returns the equator-crossing local-time drift in hours per year for an orbit that is not quite sun-synchronous (i.e., whose $\dot\Omega$ differs from $0.9856^\circ$/day). Hand-trace it for an orbit $0.5^\circ$ off the correct inclination and report the drift. (Do not run it; add # Expected output:.)
• Option C (design/your mission). If your mission (Track C, a Mars orbiter, is the natural fit) uses a sun-synchronous orbit about Mars, redo Phase 2 with Mars's $J_2 = 1.96\times10^{-3}$, $\mu_{\text{Mars}} = 4.283\times10^{4}\ \text{km}^3/\text{s}^2$, $R_{\text{Mars}} = 3{,}390\ \text{km}$, and a Mars year of $687\ \text{days}$. What inclination makes a $400\ \text{km}$ Mars orbit sun-synchronous? Record it in your MDR.

Key Takeaways

1. A sun-synchronous inclination is computed, not chosen. Demanding constant lighting fixes $\dot\Omega = 0.9856^\circ$/day, and the J2 formula returns the inclination — $98.2^\circ$ at Landsat's $705\ \text{km}$.
2. The tilt past $90^\circ$ is mandatory, because eastward precession requires $\cos i < 0$ (retrograde).
3. J2 rotates the plane, not the altitude, so sun-synchronicity is essentially free; only eccentricity, ground track, and (weakly) drag need small maintenance.
4. The orbit encodes the mission. From one inclination you can read Landsat's entire purpose: comparable lighting, global reach, and a repeating 16-day ground track — a perturbation put to work.