A personal weather website is easy to misread as one product. Current temperature, a ClientRaw parser, a credits page, and a webcam sit in the same menu, so a visitor treats them as the same kind of statement. They are not. A human sitemap is the catalog that should make the difference obvious: what was observed, what was a tool, and what was only about the operator.
The historical TNET Weather page at /sitemap.php was that catalog. It was an HTML outline of the station site, added in 2005 so people and indexers could find pages by name. This article does not republish that outline as a live directory, and it is not a second copy of the later template sitemap at /tws-sitemap.php. The useful question is how to read any weather-site map as a scientific inventory.
Historical context
Hobby weather sites of the mid-2000s grew faster than their navigation. A typical station began with current conditions and a few graphs, then added NOAA monthly HTML, a lightning page, a forecast scrape, PHP notebooks, a gallery, and a visitors gadget. The sidebar listed whatever had been finished that month. A dedicated sitemap was the first attempt to present the whole property as a structured list rather than a stack of bookmarks.
The recovered /sitemap.php page was a nested HTML list: about and location, current conditions and charts, live consoles and radar, NOAA reports, notebooks, published-data identifiers, configuration, tools, and a welcome page. Footer links on other pages pointed here as “Site Map.” That is a human map. It is not an XML sitemap for crawlers, not a generated dump of every file on disk, and not a claim that every listed URL still serves the original widget.
TNET Weather later restored this path so inbound links still resolve. Restoration of the address is not a claim that the Mesa station’s original menu is current, and it is not a live table of contents for every modern TNET product.
Three families on a weather sitemap
A useful weather sitemap sorts pages into three families. Mixing them is the usual failure.
Observations. These pages publish measurements or dated products of measurements: current conditions, charts from station files, NOAA-style monthly HTML, radar and lightning imagery the station displayed, webcam stills, and identifiers used when the station uploaded to a network. The scientific questions are units, time, location, and freshness. A sitemap entry for “Current Conditions” should be read as “here is where the station published observations,” not as a promise that the number on the page is official.
Tools. These pages are methods: a ClientRaw parser, notebooks that show a calculation, scripts, experimental parsers, PDA text views, and “how we built this” notes. A tool page may use an observation as an example. It is not itself a thermometer. Listing a parser under “Live Content” because it refreshes is a catalog error.
About. These pages are identity and policy: station location, configuration narrative, privacy, credits, site news, visitor gadgets, and contact. They answer who operated the site, what hardware was claimed, and which third-party software was credited. They do not make the temperature true. They make the temperature interpretable.
A fourth group often sneaks in: warnings and hazards. Those pages must stay labeled as documents or relays of official text. A hobby sitemap that puts “Local Hazards” next to “Charts” without a disclaimer invites a reader to treat the station as a warning service. It is not.
Why the mix matters
Search engines and humans both flatten a sitemap into a single list of URLs. The scientific cost of that flattening is category error.
Someone who lands on a parser from a sitemap link may quote its sample fields as if they were today’s Mesa climate. Someone who lands on a configuration page may treat a 2006 Davis-class station description as a current equipment advertisement. Someone who lands on a forecast scrape may call it an observation because it sat under “Local Weather” in the menu.
The sitemap is the last cheap place to prevent that. If the outline groups “Published Data” with network IDs, a later reader can see that KAZMESA12-style tokens belong with provenance, not with the guestbook. If notebooks sit under a projects heading, they stay methods. If “About Us” is not nested under “Current Conditions,” identity does not masquerade as a reading.
The same discipline applies when the modern site preserves historical URLs. A restored /sitemap.php should still teach the map, not impersonate a live station index.
How to read a PWS sitemap as a catalog
Read each line as a claim about kind, then as a path.
- Name the family before you click. “Charts” is observation display. “WD Clientraw Parser” is a tool. “Privacy Policy” is about.
- Look for live versus dated. Menus that said “Live WD Console” or “Live MesoMap” were advertising freshness. Freshness is a quality-control attribute, not a family. A stale live console is still an observation product; it is a bad one.
- Separate network products from the backyard. A sitemap that lists CWOP, a public map ID, and “dynamic banners” is listing publication channels. Those are how the observation left the site, not extra sensors.
- Treat configuration as metadata, not as a shopping list. Hardware names on a sitemap are labels for the configuration article.
- Do not promote a gadget to a measurement. Visitor maps, hit counters, and greeting banners are about traffic. They do not belong in the observation family.
- Keep official text official. Hurricane and hazard entries are pointers. The authority remains the national meteorological service.
A sitemap that cannot survive that reading is a navigation dump, not a catalog.
Distinct from the template sitemap and from XML
This URL is the human, curated map. A later page on the same host, /tws-sitemap.php, was the station-template version: a list generated from the same includes that built the sidebar, with TWS-prefixed filenames and extra project links. Two maps existed because curation and generation answer different questions. The curated map says what the operator wanted a stranger to understand. The generated map says what the template currently knew how to link.
Neither is a Sitemaps protocol XML file. An XML sitemap tells crawlers which URLs exist and when they changed. It has no observation-versus-tool semantics. Using an XML sitemap as a scientific catalog, or using this HTML page as a crawl feed, confuses two jobs that happened to share the word “sitemap.”
This article also does not replace the modern site-history hub, which is the organized index of restored historical pages in this category. /sitemap.php remains the article about how a weather site should be mapped.
Practical checklist for operators
If you still publish a personal weather site, write the human sitemap as metadata:
- Group links under Observations, Tools, and About (plus a clearly labeled Hazards pointer if you must).
- Put network IDs and file URLs with observations or with a “how to read our data” page, not with credits.
- Mark pages that require a plugin, a canvas, or a webcam as incomplete publication.
- Date the sitemap when the menu changes; a map that silently includes a retired console is a false inventory.
- Keep one curated HTML map for humans and, if you need crawler coverage, a separate XML file.
- Do not list a hit counter as evidence that the climate record is popular enough to trust.
Modern relevance
TNET Weather’s public research still depends on not mixing evidence families. Observed measurements, modelled fields, and derived summaries are different kinds of statement. A 2005 hobby sitemap that failed to separate a parser from a thermometer was an early form of the same error. The modern plain-language account of those families is how the service works. That page is the primary next step if you came here to understand how a weather-related site should be organized so a reader can tell a measurement from a method.
The restored footprint is indexed from the legacy hub and the site-history hub. Neighboring articles treat the template-generated sitemap, station configuration as metadata, and how the station published data.
This URL is not a live table of contents for Connection Outlook, and it is not a current Mesa station menu. It is the preserved address of the human map, rewritten as a catalog of kinds.