Which upcoming concerts are most likely to be heard in the neighborhoods north of the venue — from the forecast wind field and the low-frequency content of each act, fitted on 50 performance dates and 737 resident noise complaints.
Carrying north-northeast, wind from the south-southeast
Downwind
+6.3 mph
Wind
8 mph
Low freq.
Medium Pop · LC−LA 12 dB
Inversion
+3.6 °F
On the 17 comparable nights in the record: median 18 complaints, worst 117.
UnsettledThis forecast has reversed direction 2 times across the last 6 model runs. At 13 days out, wind direction is the least reliable part of any forecast.
Forecast computed 6 hours ago — Aug 23, 4:10 PM Mountain.
Beyond the forecast horizon — scheduled, but further out than weather models reach. Each moves up as it comes within sixteen days.
2024 complaint 2025 2026 Direction sound is carried Each hour of the show
Every report filed on the resident-run Ford Amphitheater noise map since September 2024, colored by season — 737 in all. That map is collected and maintained by an area resident and is not affiliated with The Monument Independent; hover any dot for the date, severity and distance recorded with it. The venue is at center; rings are one, two, three and four miles. Selecting a show above draws that night's wind: a solid arrow for the average heading, thin dashed lines for each hour of the show, and a shaded sector for the ground it covers. The arrow's length reflects how strongly the wind carries in that direction — an indication of heading, not a modeled distance. Sound travels in every direction from the venue; it simply carries farther downwind.
What this is telling you
Forecasts are available for 3 upcoming shows at the Ford Amphitheater. One is forecast at elevated or high propagation risk: Brantley Gilbert on Sat, September 5, 2026. The highest-scoring night is Brantley Gilbert on Sat, September 5, 2026, where the wind is forecast to carry sound toward the neighborhoods north of the venue at a mean downwind component of 6.3 miles per hour. On the 17 comparable nights in the record, the median was 18 complaints and the worst was 117. This forecasts complaint response, not sound level, and it ranks nights rather than predicting a count.
Last refreshed Aug 23, 2026, 4:10 PM Mountain.
Read this as a ranking, not a prediction. Held-out testing puts the correlation between forecast and outcome at +0.42 (p=0.0024). It separates busy nights from quiet ones better than chance, and no more than that. It forecasts complaints, which measure who chose to report, not sound level. Each act's low-frequency class comes from its genre, using measured spectra published by the Institute of Acoustics; assigning an act to a genre remains editorial judgment, and every call is shown on its card so it can be argued with.
How the forecast is built
Two inputs. Neither is a proxy for the other, and the pair explains more than either alone.
Downwind component
The wind vector projected onto the bearing from the venue to the center of the complaint field, 19° — roughly north-northeast. Positive means sound is carried toward the neighborhoods. Rank correlation with complaint volume across the 50 dates: +0.43 (p=0.0019). Recomputed from Air Force Academy airfield observations alone, with no weather model involved, it is +0.41 across 48 of those dates — so the finding does not rest on modeled weather.
Low-frequency content
Taken from the act's musical genre, using the four categories and measured octave-band spectra published by Fiumicelli and colleagues in the Proceedings of the Institute of Acoustics (2024). Their measure is LC−LA, the gap between C-weighted and A-weighted level, which is the standard index of bass content: dance 18 dB, urban 16, pop and rock 12. Rank correlation with complaints +0.51 (p=0.0002), and +0.40 after controlling for wind (p=0.0036). This matters because barriers do least at low frequencies — every front-of-house exceedance the venue and the city have recorded, in both measured seasons, was in the 40 Hz band.
That ordering is broadly reproduced by this venue's own record. Across the 50 dates the median submission count ran urban 21, dance 17.5, pop 8, alt and rock 3, and acoustic or orchestral 0 — though the top two categories rest on six shows and two respectively, so read the ordering rather than the spacing. Rock is not the loud category here, despite intuition — though it holds the single worst night, Godsmack in October 2024, which was also the most strongly downwind evening of that season.
How settled the forecast is
Every run is kept, so each card shows how much its forecast has been moving. What is tracked is the sign of the downwind component, because that is what the risk tiers turn on: a shift from +0.1 to +2.0 mph does not change the answer, and a shift from +0.1 to −0.1 does. A worked example from this page's own record — on 18 August the forecast for the 28th read +5.3 mph toward the neighborhoods at 9:56 a.m. and −6.0 mph away from them at 12:54 p.m., a 135-degree swing between two model cycles ten days ahead. Both numbers were the best available at the time. Neither was wrong. That is simply what a ten-day wind forecast is worth, and a page that showed only the latest figure would have concealed it.
What does not predict
Wind speed on its own (+0.05), cloud cover (-0.06) and temperature. That matters: a pleasant-evening effect, where more people are outdoors and hear more, would show up in temperature and cloud. It does not. Only the directional term moves.
The two are combined into a score from zero to four — a band for the wind, a band for the act — and each score is described by what actually happened on the comparable nights in the record rather than by a modeled number.
Score
Risk
Nights in record
Median complaints
Upper quartile
Worst
0
Low
7
0
1
8
1
Low–moderate
12
2.5
10
26
2
Moderate
9
3
7
29
3
Elevated
17
18
37
117
4
High
5
24
24
32
The five tiers are not five distinguishable steps. In the record the two middle tiers sit at medians of 2.5 and 3 and should be read as a single band; the separation that holds is between the bottom three, the elevated tier and the high tier. Note also that the elevated tier carries the worst night in the record while the high tier does not — with five nights in the top tier, that ordering should not be over-read.
Limits
Held-out testing across the 50 dates gives a rank correlation between forecast and outcome of +0.42 (p=0.0024). The method is leave-one-out: each night’s expected count is the median of the other nights sharing its score, so no night contributes to its own prediction. Nights the model called high averaged the 67th percentile of complaints for their season; nights it called low averaged the 31th. Complaint counts are normalized within season before fitting, because reporting effort drifts between seasons and raw counts are not comparable across them.
The complaint source is the Ford Amphitheater noise map, run by an area resident rather than by any government body. It is advocacy-aligned — it solicits donations for legal action against the venue — and it permits one report per exact coordinate, so its counts are a hybrid of events and locations and reflect who chose to report. It is used here because no downloadable public complaint dataset exists; the city's own records are split across several systems and are the subject of open-records requests by this newspaper. Not modeled: attendance, stage volume, day of week, the venue's own real-time level adjustments, and drifting reporting effort. Temperature inversion is shown for context but is not in the score; on its own it correlates +0.24 with complaint volume (p=0.0963), which does not reach significance, and it could not be checked against any observation — the airfield reports temperature at the surface only.
The venue's coordinate is not published in any document reviewed and was derived by trilateration from residents' own stated distances, with a root-mean-square error of 0.09 miles — adequate for the banding used here, not for precise distance claims.