We Were Asked What the Crash File Says About Darkness

What was asked, and what we did

Every crash report filed in Texas records the light condition at the scene: daylight, dark with street lighting, dark with none, dawn, dusk. That field was already in the file we pulled for Saturday's piece — 421 crashes on the corridor between Beach Access Road 1 and Park Road 22, 2016 through August 15. Nobody had asked it to sort itself by light before. Readers asked, so we sorted it.

One framing note, because it decides what the numbers mean. Counting crashes by time of day mostly measures traffic: fewer cars run at 3 a.m., so fewer crashes happen at 3 a.m. That tells you little. The useful question is not how many crashes happen after dark, but how many of the crashes that do happen end with someone dead. That is what the chart below measures.

What the sorted file shows

Darkness accounts for 27 percent of the crashes on the corridor and half of the fatal ones. Between 8 p.m. and 6 a.m. — 22 percent of the crashes — the road produces 58 percent of its deaths.

The gradient inside darkness is the part that surprised us. On stretches with street lighting, a crash after dark kills someone 2.4 percent of the time, which is close to the daylight figure of 1.7 percent. On stretches with no lighting, it is 7.0 percent. A crash on an unlit stretch of this road after dark is about four times as likely to kill as the same crash in daylight.

Geographically, every fatal crash after dark in the file is at or north of Access Road 2 — in the same unlit, undivided middle where Saturday's piece found the deaths concentrated regardless of light.

What it does not show

It does not show that lighting would have prevented those deaths. The southern miles TxDOT lit in 2023 and 2024 have had 65 crashes after dark in this ten-year window and none of them killed anyone — but that was true before the lights went in as well as after. The lit section is also the section with the signals, the lower speeds near Park Road 22 and the shortest distance to a hospital. Untangling those would take more than a crash file.

Twelve fatal crashes is a small number. One additional death in the wrong cell moves these percentages several points. Read the gaps as direction, not precision — and note that neither of the two crashes that killed three people on August 10 and August 14 happened after dark. The couple killed Friday died at 3:40 in the afternoon.

What the file does support is narrower and still useful: on this road, darkness and the absence of lighting travel with a much higher chance that a crash turns fatal, and both conditions are concentrated in the miles that have no median and no funded project.

“It is important to address Hwy 361 now.”

— State Rep. Todd Hunter, in a text message to KRIS 6 News, August 17, 2026

Why the timing matters

On Wednesday afternoon, the full Nueces County state legislative delegation — Rep. Todd Hunter, Rep. Denise Villalobos and Senators Juan "Chuy" Hinojosa and Adam Hinojosa — meets TxDOT about this highway. Rep. Villalobos told KRIS 6 News that this is not the first time the delegation has met about it.

The widening those meetings are ultimately about is expensive and distant. TxDOT's public project tracker puts the southern piece at $60 million with a 2032 letting date and the northern piece at $54 million in 2036, and programs nothing at all for the roughly eight miles in between — the miles where most of the deaths are.

Lighting is a different order of cost. The three lighting jobs TxDOT has already awarded on this corridor — one near Gulf Waters, one at La Concha Boulevard, and five miles running south toward Park Road 22 — came to about $483,000 combined. That is less than one percent of either widening piece. The argument that has kept more of it from happening is not primarily engineering; it is who pays. TxDOT has taken the position that continuous lighting inside a city is a city expense, which is why the phrase turns up in Port Aransas council workshops alongside the mayor's remark that the state ought to pay for its own highway.

We are not in a position to say what the delegation should ask for on Wednesday. We are in a position to say what is in the file, which is this: the deadliest condition on the island road, after the missing median, is the dark.

Update, August 18: the alcohol layer, and why the file undercounts it

A reader asked us to add impairment to the day-and-night view. The chart above now carries it: the shaded portion of each hour is the share of those crashes with alcohol or drugs recorded, and every fatal crash circled in the bottom panel involved an impaired driver. Adding the layer turned up two findings — one about the road, one about the record itself.

Sorted by light, the corridor's crash reports record alcohol or drugs in 4.4 percent of daylight crashes, 16.3 percent of crashes after dark on lit stretches, and 18.3 percent after dark with no lighting. Impairment shows up roughly four times as often in crashes after dark. That is thirty-three flagged crashes out of 421, so treat it as direction rather than a precise rate.

Then we checked that field against the federal fatality file, which carries actual laboratory results, and it does not hold up. Of the corridor's twelve fatal crashes in the state file, the contributing-factor field flags six as involving alcohol or drugs. The federal file, court outcomes and DPS lab work establish impairment in nine.

The clearest example is the crash that killed a 35-year-old Corpus Christi man in August 2024. The federal record puts the at-fault driver's blood alcohol at 0.200 — two and a half times the legal limit. The state crash report's contributing-factor field for that crash says only "failed to drive in single lane."

The reason is procedural, not anybody's fault: the officer's crash report is filed within days, and toxicology comes back weeks or months later. Nothing goes back to update that field. In one case here, the intoxication-manslaughter arrest came eight months after the crash.

That matters for how anyone reads the alcohol numbers above. If the file misses a third of the impaired crashes among fatalities — the crashes that get the most investigation — it certainly misses more among the four hundred non-fatal crashes, where blood is rarely drawn at all. The real share of impaired crashes on this road is higher than the file says, in daylight and after dark both. We would rather tell you the number is soft than hand you a clean-looking figure we cannot stand behind.

What the two layers say together

Impairment is in nine of these fourteen crashes. Darkness and missing lighting travel with a much higher chance that a crash turns fatal. Both are true, and neither is the whole picture.

What the fourteen rows have in common is not the driver. It is the geometry. Eight of them ended with a vehicle in the oncoming lane on a road with nothing in the middle to stop it — whether the driver was drunk, high, distracted, or simply unlucky enough to be rear-ended, as the couple killed on August 14 were, at 3:40 in the afternoon, stone sober.

That is the value of putting the layers on one page. A road cannot stop people from drinking. It can decide what happens to everyone else when they do.

How this was done

Same file, same corridor, same method as Saturday. Crashes were pulled from TxDOT's Crash Records Information System for SH 361 in Nueces County, filtered to the segment between Beach Access Road 1 and Park Road 22 using the state's own distance-from-origin measure, and grouped by the light condition recorded on each report. Dawn and dusk are counted separately from dark, which is why the panels show fourteen markers where the corridor total is twelve fatal crashes plus the two from last week that the state file has not yet ingested.

The spreadsheet and the method note are published with Saturday's piece. If you find an error in this, tell us and we will correct it in public.