Second step: fieldwork

Low-cost sensor shield improvised from two kitchen funnels and a clothes peg.

The purpose of the fieldwork was to test the computer predictions about microclimates where tender crops might be grown. We set out to do this by gathering actual temperature data from the chosen site on West Town Farm. We distributed nine temperature sensors across four points on the south-facing slope of Easterbrooks and at one point on a north-facing slope opposite. We used small, inexpensive iButton dataloggers, set to take measurements every 30 minutes, 24 hours a day.

Air temperature sensors were sheltered from direct sunlight and suspended to give a free flow of air around them. We improvised shields from plastic kitchen funnels (see above). Soil temperature sensors were wrapped in muslin and buried at a depth of approximately 15cms, marked with tent pegs.

Data was downloaded every four to six weeks and the loggers reset. Data collection commenced on 6th July 2014 and was continuous until 30th June 2015.

First step: computer predictions of microclimates at West Town Farm

Computer model of temperatures on West Town Farm (indicated by black border). Left: Frequency of days where the temperature exceeds 18˚C in West Town Farm area (darker orange indicates higher frequency). Right: Frequency of days where the temperature drops below 0˚C in West Town Farm area (darker blue indicates higher frequency)

When we began our quest to grow edible bananas in Southwest England, we identified temperature as the key obstacle to success. Frost can kill the plants and they also need a period of continuous high temperature to achieve full growth, flowering and fruiting. As a result, we decided to focus our fieldwork on air and soil temperature.

Our priority was to find a site which was unlikely to suffer severe frosts. As a tropical plant, bananas do not have a period during which they are naturally dormant and so their growing parts are not protected from low temperatures. In macroclimate terms, the Southwest is a good choice as it has relatively mild winters, particularly near the coast. However, it does not have the warmest growing season, so maintaining enough heat to produce a crop is a challenge. 

As a starting point, we generated a computer model which predicted the most banana-friendly areas on West Town Farm. The model uses the shape of the landscape to estimate the surface temperature. To do this it predicts how the shape of the land modifies the landscape; it calculates the movement of the sun across the sky, and the heating effect of the sun on different slopes; it estimates the cooling effect of wind in sheltered and exposed parts of the landscape; and it predicts the pooling of cold air at night in depressions and valley bottoms (frost hollows). 

From the maps of microclimate produced by the model we chose the slopes of a sheltered south-facing pasture, known as Easterbrooks, as our pilot site (top centre, just inside the farm boundary). This offered us the highest number of frost-free days and the highest number of days where the temperature exceeded 18˚C.

Dataloggers decommissioned

Today we collected the final batch of temperature data from our pilot banana-growing site at West Town Farm, and took down the posts and shields that held the dataloggers. Not all of the datalogger batteries lasted right to the end of the programme, but we do have almost a full year's data from the site, which we'll now evaluate and turn into graphs and charts so it's easier to see how this site compared with average temperatures for Exeter during the same period.

Overall there were seven sensors on the south-facing slope of Easterbrooks and two on a north-facing slope opposite. They were installed with the help of Dr Jonathan Bennie from Exeter University Environment & Sustainability Institute. The sensors took readings every 30 minutes, 24 hours a day from 6th July 2014 to 27th June 2015.  

Seeing changes

Looking towards Easterbrooks from the north-facing field opposite

We visited Easterbrooks today to collect the latest batch of data. We now have three rounds of data covering four months from July to October.

While south-facing Easterbrooks remains sunny and warm (if windy enough to blow a couple of the more exposed sensor housings off their posts), the north-facing slope is clearly now in deep shadow for much of the day.

In the autumn light it's easier to see the dip that our potential banana-growing hotspot sits in. It's in the upper right quadrant of the photo above. This sheltered position may also explain why the nearby trees (like those bordering the lane below) have retained their leaves.

First tests with thermal imaging camera

A normal digital image of the potential site for growing bananas at West Town Farm with thermal image alongside. White/red are hottest spots and blue/black are coolest spots.  Cattle can be seen in the distance at the top of the hill.

Button sensors

Measuring growing conditions more precisely

Temperature sensor - Easterbrooks, West Town Farm, Ide

Having selected a potential site for banana growing at West Town Farm, we're now investigating air and soil temperatures using iButton^® data loggers. The loggers themselves are tiny, but those measuring air temperatures need to be protected by radiation shields to ensure their readings aren't affected by direct sunlight. (We made these shields from ordinary kitchen funnels.)

There are seven sensors on the south-facing slope of Easterbrooks and two on a north-facing slope opposite, installed with the help of Dr Jonathan Bennie from Exeter University Environment & Sustainability Institute. The sensors will take readings every 30 minutes for up to 42 days.  We're looking forward to downloading the first batch of data and finding out what they tell us.