The reading
A weather station stands in the grove and reports every five minutes. This is its most recent observation — not a forecast for the region, and not an average of anywhere else.
Reading the station…
Connecting
One continuous line
Every hour the station has been running, drawn end to end. The warm ground beneath the line is sunlight — so the daylight on this chart is written by the daylight itself.
The record is young on purpose: it began the day the instrument did. Everything before August 3, 2026 in these grounds went unmeasured, and this page will not invent it.
On the afternoon of 4 August the chart caught a squall in full. Air temperature fell 11.9 °F in ninety minutes — 86.0° at four o'clock, 74.1° by half past five. Solar radiation collapsed from 760 to 6 W/m² five minutes before the first rain registered: the cloud's shadow arrives ahead of its water, and at five-minute resolution you can watch it happen. Total fall, 0.28 inches. An hourly regional model records that afternoon as slightly cooler than average.
Grove and model
Until this station went up, every number on this site came from a forecast model. Both readings below describe this moment. They do not agree.
Open-Meteo resolves this coast in grid cells one to eleven kilometres across, over open country, and reports wind at ten metres above the ground. The station is a single instrument standing inside a food forest, close to the plants. Where the two disagree, neither is faulty — they are describing different places. Wind and sunlight diverge furthest, which is what a canopy and a low, sheltered siting would be expected to do. Read it as a standing observation rather than a controlled one: there is one instrument here, and its exposure has not been characterised against an open reference beside it.
Four channels, still empty
Soil moisture, temperature and conductivity, on four sensors. They went into the ground on 5 August 2026. Nothing has arrived in the table yet.
They are a comparison, not a survey: one depth, four places. Channel 1 sits under closed canopy and channel 3 in open bare sand, and those two carry the argument — whether mulch under canopy holds water and steadies temperature in a way bare Florida sand does not. Channels 2 and 4 are working beds, the banana and turmeric ground and the spinach grove, which answer a different and more practical question about when they actually need water. All four at the same depth, or none of it compares. Conductivity will be published as a trend and never as a nutrient value — inexpensive EC sensors drift with temperature.
The nights nobody recorded
Two freezes crossed this grove in 2026, one in late January and one in late February. Four mango trees did not survive them. The moringa was cut to the ground and came back from its roots.
Lost · Mango TJCogshall Lemon ZestPickering
Not one minute of either night was measured here. What is known about them is what was visible afterwards, in what had blackened and what had not.
That absence is the whole argument for the instrument. On a still, clear frost night the air at plant height runs colder than the air at five feet, and cold drains downhill and settles against fence lines — so the temperature that decides which trees are alive in the morning is not the one the region reports, and it is not quite the one a station on a post reports either. It is the one measured where the leaves are.
The station has been recording since August 3, 2026. It will not save anything by itself. But the next freeze will be a record instead of a memory, and the difference between the two is whether the following winter can be planned or only survived.
How these numbers are made
Three things worth knowing about any figure on this page.
The instrument
Seven instruments in one housing, standing in the open. The anemometer is ultrasonic and the rain sensor piezoelectric, so neither has a moving part to seize or a bucket to clog. It is solar powered, which is why it stands in the open rather than under the canopy it is describing.
Every five minutes, kept
A reading is taken and written down — — observations so far, and counting. Nothing is averaged or smoothed before it is stored, which is why the record holds the squalls as well as the quiet afternoons.
What is never recorded
Conditions inside the house, and anything that would say precisely where the station stands, are discarded before a reading is stored at all. You cannot publish by accident what you never wrote down.
If the station has been silent for more than twenty minutes, this page says so rather than showing an old number as though it were current.