A new client asked me to do some math. The question posed to me: “Does this small Arizona town support the growth of pistachios?”. Like most places in Arizona, there is sunshine and water in plenty. What Arizona often lacks is the right amount of cold weather.

As days lengthen, pistachios will awake from their winter dormancy by putting out delicate baby leaves. Only if they get enough cool weather will they put energy into producing nuts.

Despite not having invented time, or calendars, plants have many ways to know the seasons. The pistachio and stone fruits use cold weather to know it’s winter. They keep track of the accumulation of cold weather to know when winter is over. Each hour of cold weather is a “chilling hour”. Different fruits need different amounts of chilling hours.

In the 1970s, crop researchers studied the pistachio to figure out how much cold weather a tree needs to nut. They determined that pistachio needs 500 hours of cool, but not cold, weather to know that winter passed. Below freezing conditions don’t count toward chilling hours. Hours above 45 degrees are too warm to count. Only cool weather between November and February counts towards the chilling hours.

Newer models treat accumulating chilling hours not as a cumulative sum, but more like interest. Below freezing is a withdrawal, and in an ideal band, you accumulate chilling ‘units’ faster. This model, the Dynamic Utah model, lives in a spreadsheet and was reverse engineered in (2004 Frankirs, et al).

The potential orchard location is at the bottom in a long narrow valley created by volcanic activity at the end of the last ice age. The mountains rise up 4000 feet above the valley floor, and winter rains can leave a dusting of snow on the peaks.

Since time immemorial, the local Yuman speaking people lived on the land. In the 1860s, Americans killed their leaders, and placed them in a concentration camp. With natives removed, and a railroad providing a connection to eastern markets, Americans developed cattle ranches in the valley. Along the river that bisects the valley, produce was grown to support the surrounding mines.

There is a good amount of historical daily weather data, but for chilling hours, I need hourly weather data.

NOAA uses satellites to measure surface temperatures. Via darksky, I can get the hourly temperature that NOAA collected. The issue with this dataset is the spatial resolution. I don’t get the area the data represents, it could be the valley floor, or the side of a canyon. Given the narrowness of the valley, and the height of the mountains, this could be an issue.

Local resident, Anges Sharpe, for whom nearby Sharpe rock is named, owns a weather station that’s been broadcasting on Weather Underground.

Let’s look at the full range of weather data.

Oof looks like Agnes plugged in the weather station after she got back from Xmas in Barbados. There were many frequent outages. After a few months, she unplugged the device for good.

The real issue with this dataset is that it’s missing the most important details, chilling hours during the pistachio ‘winter’ season.

We have two streams of data. One higher frequency with lower accuracy, and one with large gaps and higher accuracy. Sensor fusion via a Kalman filter would allow us to combine the data from the two sensors. The satellite data could fill in the gaps. The Sharpe data would ensure higher accuracy.

Before busting out the big guns, let’s check on chilling hours. Our goal is 500 hours in a season.

So in the first half of 2020, there were about 380 cumulative chilling hours. This is a fantastic start in the year. It looks like we should get enough chilling hours. What’s the difference between the sensors?

It looks like the Agnes sensor runs cooler, and there is no obvious trend. The biggest deltas are right after missing data, so they can be ignored. A couple things could be that the sensor is on a patio, or on the cooler side of the house. Or that the Sharpe home is on a north-facing hillside that elevates her higher than the valley floor. This could explain the cooler nights, and the not much warmer nights.

Comparing the temperature data, the Agnes sensor runs a couple degrees cooler. This bodes well for accumulating chilling hours.

Knowing that the satellite data isn’t too inaccurate, let’s look at the annual chilling hours. For accounting, the growing season starts in September. So the 2019 growing season runs from September 2019 to September 2020.

Year Chilling Hours
2018 429
2019 617
2020 640

At 617 hours for the 2019 growing season, this is more than enough for pistachios to shake off the winter dormancy. For 2021 growing season comes in at 640 hours. I’m expecting that this will be a good year for harvests.

So, based on chilling hours, this small valley is a good place to grow pistachios, and other fruit that need to accumulate more than 600 chilling hours.