Your first lunar investigation#
The Lab starts with a question you can change: what orbital observations intersect this place within a time window around my reference? The reference can be a real event you are studying or an analytical time of your choosing. Entering it does not make it an independently verified historical event.
1. Define the research reference#
Enter a short investigation name, latitude, east-positive longitude, and reference time. The time control is explicitly UTC; it is not interpreted in your computer’s timezone.
You can also choose Set reference on the interactive Moon and click the sphere. The resulting coordinate is a user-defined query reference on the displayed lunar sphere. It is not a surveyed landing-site position. Numeric coordinate fields remain available when the renderer or graphics hardware is unavailable.
The Chang’e 6-area starter uses an approximate location and time to make the first interaction concrete. Running it creates a new discovery. It does not reconstruct the 14/39 historical population or reproduce the historical frontier counts.
2. Choose temporal and spatial scope#
The before and after windows extend from the nominal reference time. A 30-day before window and a 60-day after window request acquisitions over those respective intervals; they are query parameters, not assertions that observations exist throughout them.
Two source-intersection modes are available:
Mode |
Meaning |
Python continuation |
|---|---|---|
Covers the query point |
The existing small point-intersection box, with a half-width of \(10^{-6}\) degrees |
Compatible with the existing Python point-query grammar |
Intersects a surrounding box |
A rectangular angular region around the reference |
Exploratory only; not silently convertible into a canonical point Study |
A box’s size is specified in degrees, not kilometres. Longitude extent does not represent a constant surface distance across latitude. See Coordinates and lunar context before treating a region as a physical measurement.
The advanced controls also accept symmetric temporal uncertainty. Acquisitions that touch the reference uncertainty interval remain overlap rather than being forced into before or after.
3. Find actual observations#
Choose Find observations. The server queries the fixed NASA Lunar ODE profile: Moon, LRO, LROC, CDRNAC4. It obtains counts before requesting metadata, retains raw response bytes, and caps retrieval rather than silently loading an unbounded archive population.
The result header distinguishes:
Count-reconciled capture: retained unique rows reconcile with the bounded count witnesses.
Partial capture: retrieval, normalization, or a budget stopped completion.
Closure not established: the available witnesses do not establish a unique complete population, for example across split longitude segments.
A successful HTTP response is not proof of archive freshness. The Lab does not currently establish an indexed-through watermark. Read Source discovery, closure, and currentness for the exact retrieval limits.
4. Inspect one object across representations#
A selected observation can be inspected through its table row, actual acquisition timestamp, and available spatial metadata. The evidence specimen shows its durable source identity, observation angles, temporal role, and a remote browse image when available.
A source center is not a footprint. An observation lacking spatial metadata is still available in the table; the Lab does not invent a polygon for it.
Table search, role filtering, sorting, and page navigation change the view. They do not automatically change the analysis population. The inclusion checkboxes and Use visible rows for analysis are explicit operations for changing that population.
5. Explore comparisons#
Once before and after observations exist, open Explore comparisons. Choose active objectives and inspect the nondominated set or a dominated pair’s actual witness. Changing plot axes changes a visual projection, not the scientific objective set.
The calculation is exploratory browser analysis. It is not a canonical Python compilation or an admitted Study. See Explore before-and-after comparisons.
6. Carry the investigation forward#
Share draft intent copies a URL containing the research definition. It does not contain the captured archive responses or guarantee that a later query will return the same rows.
Download captured investigation preserves more: the source query, response bytes and their identities, normalized exploratory observations, the included population, and any current pair result. It does not include remote browse pixels. The Python bridge can audit the response bytes offline and explicitly make a fresh point acquisition. See From exploration to reproducible computation.
When something changes or fails#
Situation |
What the Lab preserves |
|---|---|
You edit the query after a capture |
The old result remains attached to its original query; it is not relabeled as the new result |
A later request fails or is canceled |
The last reconciled capture remains inspectable |
ODE returns no records |
The result is bounded by source scope and closure; no physical-absence claim follows |
The 3D renderer fails |
Numeric inputs, discovery, tables, evidence, export, and flat context remain available |
A browse image fails |
Source identity and metadata remain; no alternate image is substituted |
The retained historical Study is accessible independently of live discovery. It remains useful during an upstream outage.