Ben's Ecology Blog

On building a worm detector

The next frontier in worm research is to be able to visualise (map) the locations of worms, in-situ, and if possible, at the paddock scale.

We know so much more about the movements of ants, spiders, beetles, bees/pollinators and the above ground invertebrates. Yet so little about everything below ground. Frankly its a blight on us all we don't know more. Okay maybe that's a bit dramatic, but for a group of organisms that are so important to soil fertility and health, we know so little about their movements specifically.

Don't get me wrong, many studies have sampled worms out in the field. So we do know about their abundance and diversity in a broad range of ecosystems, especially within agricultural landscapes.

But it's the movement of these worms, both individually and collectively that I wish to understand more. To research this we would have to make some sort of worm detector. Being able to map, not just of one species, but a community of worms in a paddock would be highly valuable data. Currently to my knowledge, there is no method able to capture such worm data.

Its a gap in the research that I am hoping with modern techniques, we may soon be able to investigate further.

My tiny lit review on worm phenomics...

The only study I found that even attempted to record/map the movements of earthworms, in-situ, both as individuals and collectively was by Mather and Christensen (1988) in Denmark. They studied Lumbricus terrestris, a widely found globally distributed earthworm species (including here in Australia). They studied this species as it regularly surfaces, especially after rain events and forms trails, making it easier to observe/record movements. To my knowledge, this is presently the closest we have got to mapping/visualising worm movements. There probably are more papers on earthworm phenomics, I just haven't had a thorough enough look.

The two authors of this study, Janice Mather and Ole Christensen, the legends that they are, then did a similar experiment 6 years later, researching the individual movements of the invasive New Zealand Flatworm, Arthurdendyus triangulatus, in the Faroe Islands, Denmark. This study differed from the former, in that in the previous study, they measured worm trails using string, but in the latter (in 1994) flatworms were placed on black plastic and their mucus trails were measured with aid of conveniently spaced water droplets.

These were the only two papers I found with any data on the distances traveled by individual worms and both still only measured and recorded above ground movements. With so many advancements in technology, along with advancements in the entire the field of phenomics since these papers, I really feel like its about time we develop some new techniques/methods.

As an aside: How far can an individual earthworm travel?

Well on the night of the 23rd and 24th of October 1987, thirty L. terrestris travelled a mean minimum distance of 9.0 m (with a s.d. 3.6 m). The distance individual worms travelled ranged from 3.8 m - 19.3 m in the study area. However, it was because the worms trails moved out of their defined study area (a dirt road) and continued into neighbouring paddocks, that made them unable to measure the full lengths of all worm trails. Further, disturbance made by tractors and footprints also damaged some worm trails. Thus, the estimates are conservative and this is why they are using this mean minimum distance value. In short, we know the average L. terrestris can travel at least 9.0m per night (n=30).

What is possible? Ideas on developing such a technology

Immediately I thought of LiDAR. Could we use LiDAR? No, probably not. But instead of sending out pulses of light (which cannot penetrate the ground), what if we use microwaves, or possibly some other wavelength of light? Microwaves is what they use in ground penetrating radar. I am currently unsure if ground penetrating radar has enough resolution to detect worms. It's difficult as we need to collect data in three dimentions, and unlike the atmosphere, the ground is very much a solid and not a gas. So maybe we need to stick probes in the ground, to capture the data.

What I imagine is this... Stick a probe in the ground that sends out pulses of a particular wavelength(s). The pulses have known strength and frequency that differ in the absence/presence of worms. How this changes as worm abundance goes from 0 worms, to 1 worm to 10 worms would require prior experimental knowledge. Thereby a probe is able to detect and record any worms nearby, along with GPS data. Do this systematically across a paddock, maybe one probe every 15m or so. Then later on the computer, chuck all this raw data into a model (with callibrations based on the soil types, and maybe other environmental variables, soil moisture/temp etc). You then match this data up spatially, with all the location data. The longitude, latitude and I would say depth, but we need a word for 'negative elevation'. Instead of being based on sealevel, it would be based on some other constant?1 Once all the data is matched up, you could generate a 3D map.

Even if we aren't able to capture exactly where worms are in three dimentions, or visualise individual worms or their burrows, could we at least create a 2D heatmap that displays the probability of worms over space, at the paddock scale? This sort of technology seems somewhat feasible in my head and would be a step in the right direction.

I am sure I am completely underestimating how difficult this really is.

Earthworm and larvae next to a partially dug up soil moisture/temperature sensor: Earthworms and larvae next to a soil sensor

But how would we know that we are detecting worms, and not other ground-dwelling invertebrates? A distinct feature of worms is their soft body. They don't have an exoskeleton - instead they have a hydrostatic skeleton! This may be a hindrance more than a benefit in detection. I am not sure, but I think the technology should try and use this attribute of worms to it's advantage. It may require preliminary studies to work out the exact worm 'fingerprint', but because worms are quite distinct to other soil fauna, I think they are a good candidate to try and detect.

We can also first try this under ideal condition (e.g, a rainy day, when all the worms have moved up in the soil profile). Would this make it easier to detect them?

Questions a worm detector may help answer...

Successfully being able to develop such a technology/methodology, would allow us to start investigating some other interesting question, such as:

Visualising worms at the paddock scale has great benefit. The magnitude of ecosystem processes/services provisioned by worms also scales with increasing area. Not only can the benefits to the soils provided by worms be better understood, the paddock scale is the scale at which farmers can manage. Knowing where worms are at in the paddock, (or at least roughly) can help a farmer be more precise with treatments/inputs into the soils. Such knowledge would help support worm populations in the short and long term.

This isn't just about the worms...

In general, this isn't just worms. We still haven't been able to crack below ground phenomics at the paddock scale. This also applies to mapping/constructing visualisations of underground roots in plants. Developing a worm detector also may help visualise root systems better. There's also the discoveries that could be made unintentionally. A 'worm detector' may discover other cool animals, like Caecilians.

Footnotes

  1. As 5m in depth changes dependent on the topography/elevation of the ground surface level - I still wonder if GNSS would work underground?↩

#o.g. obs