What have drones got to do with native plants?

Quite a lot when it comes to aerially capturing the unique "spectra" - or electromagnetic signatures - that are reflected by different plants on the ground.

This approach allows scientists to make high-tech stocktakes of the different plant ecosystems in vast areas of vegetation, such as forests and wetlands, and makes it possible to identify large-scale changes in our natural landscapes.

While all this might sound cutting edge, it's actually a commonly-used method overseas and New Zealand has been one of the few developed countries without an online "spectral library".

Advertisement

Now, the newly-opened National Spectral Library at AUT University, is expected to bring our country up to speed at a time when watching for change in our environment is today more important than ever.

Geospatial scientist Dr Barbara Bollard-Breen and Professor of Applied Ecology Steve Pointing, both of AUT, answered these questions from the Herald.

What exactly is the National Spectral Library?

Dr Bollard-Breen:

It's an online database, housed within AUT's School of Applied Sciences, and will soon be made available for public access.

The library captures plants' unique spectral light signatures, and will include native endemic species and weed species - for example, enabling us to investigate weed intrusion in our native forests.

The library will be used for remote sensing validation - for example, in analysing low altitude remote sensing data collected by UAVs, we can determine whether the data we collect is typical of that plant species.

By looking at different parts of the electromagnetic spectrum, ascertaining how each plant reflects light, and building a record of these unique markers, we can use the spectral library to identify plants on a landscape scale, across vast ecosystems or habitats.

To map a whole island species for example, we have an alternative to physically going into the field to do that.

New Zealand is one of the few developed countries that doesn't have such a library - so it's a basic tool for identifying component species of large areas of vegetation, including forests, bush and wetlands.

New Zealand also has a major skills and technology shortage in the remote sensing field - we hope the library will help facilitate and support a growing body of work in this area.
This is something that's a commonly used tool internationally, and something we wished we had had when a Masters student I supervised worked to classify the Three Kings Islands.

We couldn't physically land on the islands to validate satellite imagery; a spectral library would have enabled us to determine more about the species on the Three Kings Islands.

Can you explain what "spectra" is - and why each plant has a different signature?

Dr Bollard-Breen:

Spectra are the unique electromagnetic signatures reflected by plants.

Each plant species reflects differently on the electromagnetic spectrum, creating a unique spectra.

Plants reflect light differently as a result of factors such as the shape, size and orientation of its leaves, the amount of chlorophyll in its leaves, the presence of other chemicals in its cells, and the cell arrangement and airspaces in the plant's leaves.

These many factors give the plant's light reflection a unique signature that we can't always detect with the naked eye.

How do you go about measuring or recording it?

Dr Bollard-Breen:

In the field, we use a handheld device called a spectradiometer to measure the light reflected by the plant, through multiple readings.

This hyperspectral imaging measures tiny bands of the electromagnetic spectrum and gives us a spectral curve - a visual profile of the colour characteristics of that plant's light reflection in the electromagnetic spectrum.

We measure at two-nanometre bands, giving a much higher resolution image than the 100 nanometre bands satellites typically capture.

These measurements are used to record the species' spectral reflectance signature, which can then be used to automatically identify different plants in future surveys of large areas - similar to fingerprinting - enabling scientists to validate their plant species classifications.

We also collect voucher specimens of the plants we measure, which are lodged in Auckland Museum's Herbarium - another important means of enabling the verification of plant data, and also adding to the plant specimens preserved for use by future generations of botanists.

We also use unmanned aerial vehicles (UAVs) or satellites to capture information about large areas.

UAVs are excellent tools to have at our disposal as they allow us to capture much higher resolution footage than a satellite can deliver.

With lots of cloud cover in New Zealand, visibility can also be a problem with satellite footage; with UAVs flying at a much lower altitude than satellites, we can get clear images more of the time.

With UAVs, we take still shots every one to two seconds while flying at low altitude, using both a multispectral camera and a high resolution colour camera.

We can also capture video footage.

We then stitch these images together using specialist software, creating an aerial mosaic of the area and enabling us to analyse the ecosystem we've surveyed.

We use one of two types of UAV - fixed-wing, for large-scale surveys, or rotary, for smaller areas requiring less battery time.

Can you tell us about the work that has already been carried out in Antarctica?

Professor Pointing:

Using a fixed-wing UAV, we were able to measure the extent of cyanobacterial mats in the McMurdo Dry Valleys region of Antarctica - a region designated by international treaty as an Antarctic Special Managed Area, reflecting its unique and threatened status.

These cyanobacteria are the dominant life in this threatened part of Antarctica, and using remote sensing via UAV means we can survey them and how they respond to climate change and disturbance, with minimal harm compared to on-foot surveys.

This kind of low-altitude remote sensing using UAV had never been done before in the Dry Valleys, but this pioneering work by AUT drone pilots has resulted in a code of best practice for future use of UAV by other researchers.

We now have multispectral images and hyperspectral curves for these cyanobacteria species, new information that can be used in the future to detect differences between species of cyanobacteria in Antarctica and how they respond to change.

It is important to differentiate between species because they often perform different roles - for example some species are able to introduce vital nitrogen - an essential nutrient for cyanobacterial growth - to the ecosystem by fixing it from the atmosphere, while others cannot.

Using UAVs facilitated great research advances, enabling us to survey more extensively and to gain a better understanding of the Dry Valley region, as we wouldn't physically have been able to get to all the places the UAV covered each season.

It is cost-effective field science.

How is New Zealand performing in this branch of science?

Dr Bollard-Breen:

As we don't have a spectral library, New Zealand has some catching up to do in this area.

New Zealand has some excellent remote sensing specialists, however we unfortunately have too few of them - geospatial science is on New Zealand's skills shortage list and we have a major lack of capacity and technology in this important area.

AUT hopes to facilitate gains in this area, by putting tools such as the National Spectral Library in place, training skilled geospatial scientists, collaborating with industry partners such as Auckland Museum and others, and supporting the research efforts of individuals and institutions around the country.

How important and valuable will this research be in the context of climate change?

Dr Bollard-Breen:

The National Spectral Library will allow New Zealand researchers to map the country's native and invasive vegetation, and in high resolution - essentially capturing a high resolution time series depiction of the country's plant ecosystem.

This information will be invaluable to assessing the impact of climate change.
It will allow us to identify changes over time - including alterations in the health, abundance and variety of vegetation - and will help us understand how different plant species react in different situations.

We can also focus on key regions we believe are under threat, such as Antarctica, to accelerate our understanding of those areas and possible future interventions.