Project Carbdown: A Review of Year 1

Learnings from one year of Enhanced Weathering field experiments

Project Carbdown, our science project aimed at measuring the CO₂ sequestration by enhanced weathering with rock dust on croplands, has been going on for a year now. It is time to look at the results of the first year, draw some conclusions and adapt the ongoing experiments accordingly.

After a short introduction this article summarizes our learnings and conclusions. These will lead to some changes to our ongoing experiments and have also triggered the development of a new, additional experiment, the “Carbdown XXL-Lysimeter Experiment” that we will announce in a few days.

Welcome to the scientific process!

“Measuring the slow dissolution of 4 kg of basaltic rock reacting with CO₂ on one square meter by looking for reaction products in various forms, either dissolved in liquids or in solid form, which try to escape our view into several directions (into soil, groundwater, or plants) and by combining themselves with other stuff is really hard, since we are looking for only about 40 grams per year hiding in a ton of soil.”
Dirk Paessler, founder and CEO @ Carbon Drawdown Initiative

“After one year of enhanced weathering field experiments, I feel like I did upon completion of my PhD: I now know how little I knew at the start – and how much more complex it all is than I assumed.​”
Dr. Ingrid Smet, geologist @ Fieldcode, team member Project Carbdown

A Short History of Project Carbdown

In October 2020 our CEO Dirk Paessler met Prof. Jelle Bijma (Alfred Wegener Institute) for the first time. Jelle wanted to do experiments with olivine on croplands, but it would still take several months or even longer to get some funding, which would mean he would miss the next growing season and effectively lose a whole year. A year that we do not have, the climate crisis is urgent. 

In a matter of 14 days the two entrepreneurs Dirk Paessler and Ralph Koczwara decided that they will fund and support the project to speed up the scientific process. The project team grew further and Prof. Jens Hartmann (University of Hamburg), Dr. Mathilde Hagens (Wageningen University), Dr. Ingrid Smet (Fieldcode) and Ralf Steffens (Carbon Drawdown Initiative) joined the core project team. A little later Dr. Elisabete Trindade Pedrosa (AWI), Dr. Maria-Elena Vorrath (AWI), Dr. Marcel Hoosbeek (Wageningen University) as well as Lukas Rieder, Xuming Li and Giovanni Galeasso (all University of Hamburg) also joined. By the way: Most of the team has not to this day met in person due to the pandemic, even after one year of working together. We are finally planning a summer meeting.

Finally, in January 2021, we publicly announced “Project Carbdown”, our attempt to measure the weathering rate of rock dust on croplands, a method of generating negative emissions to mitigate climate change. 

We wrote: “At the end of 2021 our goal is to be able to tell how much CO₂ has been removed from the atmosphere per m² in one year. We want to be able to assess possible side effects and come up with a simplified and easy to use monitoring concept for the CO₂ removal (what metrics will work well?). We then plan to test further in 2022.”

The project consists of an interconnected series of experiments: We run lab-based experiments in Hamburg and Bremerhaven, cropland-based experiments in nature in Wageningen (NL), Bramstedt, Fürth (DE) and Larissa (GR) as well as extensive analyses in several laboratories - all with synchronized soil, rock dust, instruments and methods. The chemical/physical analyses are accompanied with hundreds of electronic in-situ sensors that take measurements every 20-60 minutes in the fields. More than 20 people are involved and many of them work pro-bono for this project.

Here is a photo album of our work from the first year and there are also several videos and TV programs about Project Carbdown (links at the end of this article). A good introduction to the topic EW is this background article in Spektrum der Wissenschaft (German) about basalt based negative emissions.

On our test fields the basalt was applied manually.

The Challenge

We are throwing 4 kg of basalt on 1 m² and expect 1% of that (40 g!) to do something over a year. Then we try to measure the reaction products of these 40 g through a thick curtain of soil-heterogeneity, bio-effects and measurement-variations. Is this possible at all?

Our ultimate goal is to prove e.g. that we can measure the minimum amount of weathering process by measuring the reaction products that were released from the weathering of the basalt/dunite, preferably measured with cheap electronic sensors in the ground.

Enhanced weathering will only succeed as a climate relevant method for negative emissions if there are reliable ways of measuring the drawdown of carbon from ambient air. A key booster for this scientific process would be to find a quick and cheap measuring concept that can easily be rolled out and cheaply scaled so that we can quickly learn how to create a good environment for the weathering process (avoid slow down effects, e.g. lack of water) by doing lots (hundreds, thousands?) of field experiments. The current lab-based way of monitoring this is not practical for such a process.

Our First Results (Short Version)

Before we dive into the details let us try to summarize what we have learned so far:

Learning #1: It actually works!

In our closed-box lab experiments (soil column experiments by our team member Prof. Jens Hartmann, University of Hamburg), the weathering effect can be clearly measured. We are seeing signals for the CO₂ collection in the order of 0.5%-1.5% per year of the basalt’s total lifetime CO₂ capturing potential of about 0.3-0.5 tons of CO₂ per ton of basalt. 

At a rate of 1% per year it would take about 70 years to use half of the basalt’s CO₂ collection potential: That would be too slow to stop global warming!

But these experiments use ambient air (420 ppm of CO₂) and soil samples without plants and with less natural bio-activity (fungi, bacteria, fauna, flora), which we expect to enhance the weathering rate. Beerling at al 2020 estimates an 1-6 enhancement factor of weathering by biological processes (does not provide actual measurements though). We expect such an enhancement because in the soil there is more CO₂ (10x more in central Europe, up to 100x in tropics) than in the ambient air and because fauna and flora do things that help (we think, also see Vicca et al 2022).

Conclusion: As many studies did before we can also show that enhanced weathering actually works. 

Learning #2: For science use more rock dust!

With our outdoor field experiments in “open nature” it is - as expected - much harder to separate the signal (in our case, measuring the effects and chemical products of enhanced weathering in soil/water) and the noise (everything nature is doing all the time, including weather, plants, etc.).

We might have been too optimistic with our initial decision to use 40 t/ha (=4 kg/m²) of powdered rock in our field experiments: Assuming the annual rate is in the order of 1-2%/year this means it is only 40 g of “stuff” per m² per year which we try to find in about a ton of soil. So this might generate a signal too small to compete against nature’s noise. What makes it even worse is the fact that some of the weathering products seem to be sucked up by soil and biogeochemical processes. Our current thesis is that a certain threshold in the amount of weathering products needs to be overcome until the weathering products become fully visible and this threshold depends on the type of soil. 

This does not mean that the weathering is not happening at all, it clearly is, but with only little basalt it becomes very hard to measure the weathering effect outdoors. It seems to make complicated isotope analyses or similarly complex chemical analyzes necessary, which slows down the scientific process too much. Our initial isotope analysis already shows promising results. 

With our goal in mind to better understand enhanced weathering and to develop it into a climate-relevant method of negative emissions we will in the future need hundreds or thousands of field experiments (with variations of e.g. soils, crops, climate, rock-type, etc.) with fast turn-around-times. This requires a fast, easy-to-use and cheap way of measuring (= without a lab). 

But: On one test plot with much more powdered rock (400 t/ha) we saw that the signal is already much clearer even without the more complex measurements.

Conclusion: We will apply more rock dust to our fields in 2022 and we will do this in several steps: 100, 200 and 400 tons per hectare (instead of the 40 t/ha last year) so we can develop a characteristic curve to estimate the weathering amounts for smaller rock dust amounts.

Learning #3: Supply enough water! But not too much!

Our lab experiments show that the weathering rate depends heavily on the amount of water that is available. It looks like the best place for weathering are the tropics (>1000mm of rain and high temperatures), ideally on old, weathered soils so the newly added rock dust can be quickly weathered.

But not only the weathering process requires water -- also measuring the weathering products chemically requires pore/ground-water, and that was a problem: On two of our fields there was not enough rainfall so that we could not consistently collect water samples from our lysimeters and rhizons every 2 weeks. In Fürth we had 5 months without chemical data (just electronic data). In Bramstedt on the other hand we had 3 months of too much water in winter: The field wasn’t drained enough, we couldn’t even walk on the field and some sensors were killed by the water.

These problems combined with the too-small-signal made the analyses really complicated.

Conclusion: For our experiments in 2022 we will apply controlled irrigation wherever possible.

Learning #4: We need an experiment somewhere between lab and nature

Our experiments have been on two extreme ends of the spectrum: On one side we do lab-based experiments almost without involvement of nature, flora and fauna. On the other side our field  experiments are fully embedded in nature. In the lab we miss the effects of plants, fungi, bacteria and all the crawly soil folks while on the fields all these natural effects made measuring the weathering almost impossible.

The solution is to set up so-called mesocosms: A mesocosm “is any outdoor experimental system that examines the natural environment under controlled conditions. In this way mesocosm studies provide a link between field surveys and highly controlled laboratory experiments” (Wikipedia). In soil science these buckets are called “lysimeters” and we decided to build 20 large XXL-lysimeters to provide our link between the lab and nature..

Conclusion: We will build a new experiment with a set of 20 large custom-build XXL-lysimeters. Announcement coming soon, stay tuned!

Learning #5: It takes 4-8 weeks until the soil “heals”

In our electronic and chemical data we could see huge spikes in the first weeks, only after 4-8 weeks the data from different sensors and sampling points showed consistent results. Obviously it took so long until the soil, which we disturbed to put our instruments in, had settled down and the whole system worked normal again. More water seems to make the settling faster.

Actually our one-year-old-fields have become a valuable asset now: The instruments and sensors have been in the ground long enough and we have good historical data. If we now apply additional (and much more) rock dust, we should be able to observe the weathering process pretty well. And when we build new experiments we will wait with rock dust application for a few weeks.

Conclusion: Build new experiments first, let them settle, then apply rock dust.

Learning #6: Finding a real-life sensor-based proxy metric for EW is hard

Along with the chemical lab metrics (based on soil/water samples) we have put hundreds of sensors into our test fields that send monitoring data every 20-60 minutes. This data is very helpful in understanding the constantly changing situation in the ground and helps to fill the gaps between the sample data which is only available every 1-4 weeks.

But: From the data that we have gained until now we come to the conclusion that sensors for pH and electrical conductivity may not be suitable to track the weathering in soils when an amount like 40t/ha of rock dust, which would be a rate we consider to be achievable for large scale rock-farming in the future, is used (we will try the same with more rock dust in 2022). Compared to other things going on in the soils the chemical weathering products can not be properly tracked with these cheap, off-the-shelf sensors when only “normal-for-agriculture” amounts of rock dust are used.

Conclusion: In order to measure weathering in real-life scenarios, other tracing elements together with the respective sensors need to be considered/developed.

Other Learnings

Apart from these bigger learnings we also had a few other aha-moments:

  • On various occasions we had massively different measurements, from the sensors and also from the chemistry, on experiment plots that had similar treatments but that were located in different parts of the field. We underestimated the heterogeneity of the soil and the underground structures/horizons. At least with “only” 40 t/ha of basalt these disturbances had such a strong effect that it is hard to understand the meaning of some of our numeric data. This does not mean that the weathering has not happened, we just were not able to measure the small signal. This fact will become a major obstacle for any EW carbon accounting project that aims to go beyond working with estimates/models and use in-situ measurements to account for carbon more accurately.

  • One easily underestimates the strong effect of temperature on the weathering rate. Every 10°C less in soil temperature can slow down the weathering rate by an order of magnitude (see Pogge von Strandman et al. 2022). When comparing lab results at room temperature with results from outdoor experiments at Northern latitudes with cold winters this needs to be taken into account.

  • During the year, we also looked at other rock dusts that could potentially be used and found that some rock dusts are not as good for EW as you would expect from their name or chemical composition. A mineralogical analysis of the rock dust is mandatory to find out if it is at all useful for weathering.

  • In this context there is one important aspect: there needs to be consensus on the way we calculate the EW potential of a type of rock. Currently, there are multiple approaches and we look forward to and would like to encourage further developments of this field of science.

  • Another scientific problem: We need to find an ‘inert’ trace element specific for each rock material to monitor the process in the soil water chemistry. We used alkalinity, ion-balance, isotopes and DIC/DOC to track the weathering products, but focussing on specific major and/or trace elements like Mg (but that’s also taken up by plants), Sr or Li (depending on the rock’s element content) could be more helpful (e.g. we found that some soils “sucked up” a lot of the cations from the weathering which causes our measurements to likely underestimate the actual effects). The best tracer element would be one which (1) the plant or soil organisms do not want to do anything with AND (2) is present in sufficient concentration in the rock dust so that it can be measured when released during weathering AND (3) can be measured live and in-situ. This would likely require building specific sensors for monitoring.

  • When we analyzed the plants from the cotton field in Greece we found that the application of 40 t/ha of olivine rich rock dust did not seem to have had any effect, positive or negative, on the cotton crop yield. Also the nutrients within the plants at flowering stage were within the same range as for those cotton plants without any olivine treatment.

  • Applying 40 tons of rock dust per hectare (4kg/m²) at farming scale can be a challenge. It is a lot of stuff, the application process can be quite dusty (surprise!), it can be easily blown away (potential health hazard) and without proper equipment the farmer may need to drive over the field too often (soil compaction). We likely need to convert the dust into dissolvable granules for spreading at large scale in the future.

  • In order for enhanced rock weathering to be globally applied as a CDR option in combination with agriculture, it seems that in a first order theoretical/mathematical/model CO₂ rates will be applied depending on the type of rock, the type of soil and the climate. 

  • Such simplified models would need to apply large safety buffers to the calculation of the sequestered CO₂ (based on lab/model data) to become widely accepted, but this could underestimate the amounts of actually sequestered CO₂ in the open field without in-situ measurements. Which in essence is what we try to develop with Project Carbdown.

Our First Results (Long, Scientific Version)

If you are interested in even more details we have a few additional documents for you (and we might add some more in the next weeks).

Videos and TV Programs About Project Carbdown

If you want to learn more about Project Carbdown, here are a few links for you:

Dirk Paessler