A study by Hutchinsons-sponsored PhD student Sarah Barnsley has revealed it is possible to use multispectral images collected from a manned aircraft to assess the availability of important flower species within hedgerow and wildflower margins.

Being able to make such assessments remotely could be much quicker and more cost-effective than commissioning a full botanical survey on the ground, yet still provide valuable information to help manage resources that enhance the biodiversity and productivity of arable enterprises, says Ms Barnsley.

Pollinators are really important from an arable perspective.

Three-quarters of crops globally rely on pollinators, yet in Great Britain there was a 33% decline in numbers between 1980 and 2013.

Increasing the proportion of flower and nectar availability is a relatively simple measure, which means, at the very least, we can prevent lack of food being a factor limiting the pollinator community.

The research, conducted at the National Helix Technology Development Farm in Northamptonshire, involved taking aerial images in March, May and July 2019, to coincide with the flowering periods of five key species: blackthorn, hawthorn, red campion, bramble, and common knapweed.

Ms Barnsley says: We focused on early spring, because that is a time of year known to be particularly poor for nectar availability for pollinators.

June is also a nectar gap.

Two image resolutions were tested 7cm and 3cm per pixel with all results ‘ground truthed against field assessments.

Flowering plants At both resolutions the research found areas of flowering plants were clearly visible and, in some cases, it was possible to distinguish individual species, notably blackthorn and bramble.

Identifying other species proved more challenging though, especially where flowering occurred at the same time, with species sharing a similar appearance or colour.

Hawthorn accuracy, for example, was lower because cow parsley was flowering in the margins at the same time.

Follow-up research is therefore building on this first phase of the project by investigating which flower species superficially appearing as similar colours could be classified separately from one another and assess how the area classified as a particular flower species translates into numbers of flowers on the ground.

Within this, drone imagery with 3cm resolution has been trialled to see if oxeye daisy and common knapweed flowers can be accurately distinguished from control species with similar flower colours, namely yarrow and rose bay willow herb.

The final report findings are due to be submitted for publication later this year before the next steps for the project are defined.

Ms Barnsley says research such as this could theoretically pave the way to one day developing digital mapping systems that can accurately identify the type and approximate number of key nectar-rich species in any given area.

She acknowledges that mapping to this level of detail is still some way off, although aerial imaging, particularly higher resolution satellite imagery, is developing very quickly.

Satellite images California-based Planet satellite labs, for example, is already gathering daily satellite images to 50cm resolution, and even finer resolutions could become publicly available in the next few years, she says.

Were starting to head towards resolutions from satellite imagery that would be feasible to do this kind of work over larger areas at a sensible cost.

Accuracy levels are not quite there yet, but I do think that in the next five to 10 years they probably will be In the meantime, Ms Barnsley believes simpler assessments of flower availability still offer very valuable information for farmers.

One option being investigated is to group flower species by colour, giving a simple measure of the diversity of flower and pollinator communities and potentially allowing the abundance of pollinator species to be estimated, she explains.

Flowers develop different traits that attract different pollinators, so a diversity of flower colours is potentially providing a diversity of resources for different species in the pollinator community.

Hoverflies, for example, like yellow flowers, and bees are known to like purple flowers.

Previous research has shown that across Britain, 22 flowering plant species provide 90% of the nectar supply.

Hutchinsons environmental services specialist, Hannah Joy, says: If you know there are a lot of flowers, there is a good chance youll be supporting many of those 22 species.

The research may not yet be absolute, but it doesnt need to be.

Indeed, resampling imagery from phase one of the project to lower 10cm and 15cm resolutions suggests it is still possible to accurately identify clusters of similarly coloured flowers on the ground, potentially providing a more cost-effective way of delivering useful information for farmers in the future, Ms Barnsley adds.

Monitoring images taken at different stages through the season and over multiple years will help identify gaps in pollinator resources across the farm and show how existing resources are changing over time.

This may help decisions, such as when to reseed flower-rich margins, where to place new stewardship options, or help tailor species selection to fill gaps in nectar availability early or late in the season.

Ms Joy says: If you want to make the biggest impact on your farm, you could target measures where there is currently nothing.

While the current research has focused on pollinator resources, Ms Barnsley says mapping such as this could be equally useful in the future for assessing other biodiversity groups, such as species that support beneficial insects in field margins, or trees that provide food for birds through the year.

Pollinators Ms Barnsley says: If you increase flower abundance, thats good for pollinators and for crop pest predators and other wildlife.

If growers or land managers are already using crop monitoring via drone or airborne imagery, then it makes sense to extend this into the adjacent habitat to see what is already being provided.

This will help farmers make management decisions around how to improve what is available at particular times of the year.

The research is still at an early stage, but its a starting point that sets a really good foundation to show what we can potentially achieve.