Site characterization

Scope of the site

The compilation of sites where spreading and monitoring occurs is called the project area. Each project registered with Rainbow shall have one project area which is described in the initial Site Characterization Report contained in the validated PDD, plus any additional appendix Site Characterization Reports when new sites are added for successive spreading events.

The project area is composed of sites which will be statistically analyzed together with similar:

• Administrative oversight and jurisdiction

• Geographic area

• Timing of spreading (less than 5 years between consecutive spreading events)

The depth of the NFZ should be at least 20 centimeters, or the tillage depth plus a 10 centimeter buffer, whichever is deeper. Deeper or shallower NFZs may be considered on a case by case basis if the Project Developer provides sufficient justification based on site hydrology or agronomic practices, including but not limited to the depth of the water table, depth of plant roots, or disturbance from installing sensors in non-tilled soil layers.

For advancement of ERW scientific understanding, it is strongly recommended that on a small representative subset of the deployment area, samples are taken beyond the depth of the NFZ (60-100 centimeters, or the end of a relevant soil horizon), in order to research inner-NFZ transformations such as secondary mineral precipitation and cation sorption.

Site characterization report

Project Developers shall submit a Site Characterization Report with the PDD during project validation, before any rock spreading occurs, that:

• justifies why the project area is appropriate for ERW under baseline conditions before any rock spreading occurs, and is expected to lead to measurable CDR

• demonstrates that spreading feedstock in the project area meets the Environmental and Social Do No Harm requirements, is not expected to cause adverse effects on the environment, and is not upstream of peatlands

• delineates the spatial extent and of the treatment and control plots

• describes any stratification approach used and the stratification results

• details the number, location and management strategy of control plots

• describes the site sampling plan (number of samples to take and sampling approach)

• provides results and sources for all components listed in Table 1 for the site, soil and FFZ.

The characteristics listed in Table 1 shall be reported in the Site Characterization Report for each site in the project area. Sources may include measurements from , data from farmers, secondary databases and soil maps, remote sensing, or other well documented and reliable sources.

These characteristics shall be used to justify the modeling estimates, the control/treatment structure, stratification, and sampling protocol (e.g. where to install measurement infrastructure, at what depth, frequency of measurements...).

Table 1 The characteristics outlined in the table shall be included in the Site Characterization Report, prepared by the Project Developer for project validation.

Site

The following characteristics apply to the Near Field Zone (NFZ).

Some agronomic characteristics listed below cannot be reliably projected multiple years in advance, because farmers may make such decisions annually. Project Developers should provide their best estimates of agronomic practices upon project validation, and actual practices shall be documented and updated throughout the crediting period.

Characteristic	Description
GPS coordinates	GPS coordinates and map of extent of site
Cropping system	Past and expected crop types and rotation schedule
Root depth	Maximum root depth of expected crops
Fertilization practice	Frequency and amount of agronomic pH control and fertilizer use in previous 3 years, and expected use during the crediting period
Tilling practices	Frequency and depth of tilling practices in previous 3 years, and expected use during the crediting period
Irrigation practices	Frequency and amount of irrigation in previous 3 years, and expected use during the crediting period
Local climate	Temperature, rainfall, humidity, annual and monthly average
Soil horizons	Depth and types of different soil horizons, especially changes related to vertical infiltration of water (e.g. hardpans, plow pans, caliche layers...)

Soil

All items below shall include measurements of both the baseline composition and the spatial and temporal variability. They shall include averages and distributions (standard deviation, variance...). Those with an asterisk (*) are optional.

The following characteristics apply to the Near Field Zone (NFZ).

Characteristic	Description
Soil type according to World Reference Base for Soil Resources (WRB)	Indicates many other soil characteristics and the particle size distribution and which affects water retention and mineral dissolution rates
Soil and porewater pH and buffer pH	Determines the acidity of the environment, as lower pH enhances mineral dissolution and CO₂ capture potential.
Base saturation	Reflects the proportion of exchangeable base cations (e.g., Ca²⁺, Mg²⁺) available, which influences soil's ability to neutralize acidity and facilitate weathering.
Bulk density	Determines soil compaction, influencing water infiltration and the interaction of crushed rock particles with soil and water.
Mineralogy	Identifies existing weatherable mineral concentrations in the soil, determining baseline weathering
Heavy metal content	Characterize the starting conditions of the soil, assess the risk of exceeding safe pollutant thresholds
Cation exchange capacity (CEC)	Measures the soil's ability to hold and exchange cations, affecting the retention of weathering products like Ca$^{2+}$ and Mg$^{2+}$
Total alkalinity	Indicates the baseline ability of soil and porewater to neutralize acids and CO$_2$ sequestration capacity
Porewater DIC concentration	Represents the baseline dissolved inorganic carbon in the NFZ, used for tracking CO₂ capture and movement through the NFZ
Soil organic carbon concentration	Influences microbial activity and CO₂ production or storage, and is an important supplementary indicator to measure (for co-benefits or ESDNH)
Soil inorganic carbon concentration*	Baseline measurement of pre-existing carbonates, important for assessing the net carbon removal potential of ERW

FFZ

The following characteristics apply to the Far Field Zone (FFZ).

Characteristic	Description
Hydrologic flow paths	Expected water flow paths, expected near-term carbon fate (groundwater, watershed, rivers, ocean entry point, aquifer), potential horizontal movement
Final carbon storage reservoir	Carbonate minerals, DIC in groundwater, DIC in oceans
Groundwater system durability	If "Final carbon storage reservoir" is DIC in groundwater, the fluids passing through the groundwater enter a flow path with high durability
CDR in vadose zone	What processes could cause CDR increases or losses below the depth of the NFZ (e.g. secondary precipitates)
Water table depth	Average and variability across deployment area
Surface water organic carbon	Sites must not have expected substantial organic carbon destabilization in downstream surface waters, e.g. from being upstream of peatlands

Project Developers shall describe the sampling protocol used for taking pilot measurements and justify why it was appropriate. This description must consider the number of sub-samples and composite samples, location of sampling, timing of sampling, secondary sources used, compositing and homogenization steps, and statistical analyses done on samples (e.g. average and distribution).

Stratification

Project Developers should perform to group the plots within a deployment area according to their key characteristics that influence CDR.

Stratification must combine key climate and soil properties to delineate strata that are relatively homogeneous in factors influencing ERW. The stratification approach shall be developed during project validation, prior to rock spreading, although this may be revised at later auditing events. The purpose is to:

• designate treatment and control plots per strata (at least 1 control plot per strata), reducing variability and improving representativeness of control plots, and

• (optional) used for spatial extrapolation.

Stratification shall be done using evidence from, data from farmers, secondary databases, remote sensing, or other well documented and reliable sources.

It is recommended that Project Developers perform this step in GIS software as a multi-criteria analysis that overlays layers representing each characteristic, but other approaches, methods and software can be considered on a case-by-case basis.

There is no minimum requirement for the number of strata in a project. For small or homogeneous project areas, the entire area can be grouped into a single stratum, with one representative control plot assigned to it.

Similarly, there is no minimum number of strata or variance level within strata, so Project Developers may choose to effectively skip this step, by grouping all fields into one loosely-defined heterogeneous strata.

Statistical methods and grouping variables

The suggested climate and soil properties to delineate strata are listed below. The properties in bold are strongly recommended for stratification, but Project Developers are encouraged to use as many properties as reasonably possible to increase the likelihood of statistical significance. Additional properties not listed here may be considered if justified as relevant.

• Soil or porewater pH

• Soil type

• Feedstock application rate and timing

• Feedstock type

• Type of crop grown

• Soil moisture

• Soil texture

• Topographic Wetness Index (TWI)

• Temperature

• Precipitation

• Slope

• Wind exposure

• Cation Exchange Capacity (CEC)

• Soil organic carbon

Categorical variables (e.g. soil type) should be treated as separate values, unless the Project Developer can justify why multiple types should be combined.

Continuous variables (e.g. pH, soil bulk density...) should be grouped using one of the approaches listed below, and justified by the Project Developer:

• Standard deviation intervals

• Equal intervals

• Quantiles

• Domain-specific thresholds (e.g. acidic, neutral and alkaline soils)

• Natural breaks/gaps

• Statistical clustering models (e.g. hierarchical or K-means clustering)

The Project Developer shall provide a statistical approach and results for grouping plots that are sufficiently similar to one another into strata.

Steps and implementation

The steps include:

• Define the variables for stratification within the project area (see minimum required variables above)

• Statistically assess the variables to establish strata and . Summarize the cutoff values for criteria.

• Provide a summary report listing each stratum, its value for each variables, and names, GPS coordinates and map outline of each parcel.

• Assign at least one treatment and control plot to each stratum, ensuring representativeness and minimizing bias (see more details on control plots below).

• (Optional) Split the treatment plot into a high-density sampling and low-density sampling plot (see more details on treatment plots below).

The final stratification approach shall be described in the PDD and include:

• a GIS generated map showing the extent of the total project area and the location of different strata.

• list of the different strata types identified, with the number of disparate strata plots and the total area covered by each strata.

The stratification setup may be revised and plots reassigned to different strata throughout the project's crediting period if:

• new sites are added,

• successive spreading events alter the representativeness of sites, or

• as a result of general learning and improvements.

This may occur if, for example, initial measurements show that a given treatment plot is more similar to a control plot in a different stratum. Such revisions should be submitted to Rainbow for approval before implementing a MRV approach that differ from the originally validated Monitoring Plan and Sampling Plan. Any revised stratification approach shall meet the original requirements of site setup (e.g. adding plots to a stratum would increase it's total area, and may require additional control plot area).

Plot types

Results from sampling and validation-stage measurements (before any spreading occurs) shall be used to designate treatment and control plots within the project area.

Treatment plots

Treatment plots are the areas where Project Developers have spread feedstock. Within each stratum, the treatment plot may be divided into high-density sampling (HD) and low-density sampling (LD) treatment plots.

High-density (HD) sampling plots

• small, intensively sampled fields

• used to generate the main NFZ measurement dataset

• provides detailed, high-quality measurements and robust datasets

• spatial extent must match the size of their corresponding control plot, depending on the total stratum size (as detailed below).

Low-density (LD) sampling plots

• larger areas with reduced sampling frequency

• used to scale ERW deployments

• measurements from LD plots

  • provide a validation check, ensuring results are within the same range of the corresponding HD plot, and

  • are added to the HD plot's NFZ result dataset to estimate total treatment plot CDR

Project Developers must collect enough samples in LD treatment plots to ensure results are sufficiently similar to those observed in HD treatment plots. Project Developers shall describe in the Sampling Plan in the PDD the statistical test/s they plan to use to assess similarity of results (e.g. t-tests, Mann-Whitney tests, ranges of standard deviations, overlapping confidence intervals...).

The minimum sampling density for LD treatment plots is 1 soil sample per 15 ha (0.0667 samples/ha) and 1 aqueous sample per 45 ha (0.022 samples/ha), or 3 samples per plot of aqueous and/or soil samples, whichever is larger (this ensures that even very small plots will have multiple samples and generate a distribution of results). Sampling density requirements for the HD treatment plot are described in the Sampling and measurements section.

Alternatively, Project Developers may opt to maintain only HD treatment plots.

Control plots

Control plots are used to measure baseline weathering and CDR in the NFZ that would have occurred without the project intervention. This is subtracted from the treatment plot's CDR, to only issue credits for CDR that occurs beyond business as usual.

Control plots shall be selected to be representative of their corresponding treatment plot and avoid contamination of weathering material from treatment plots. Each control plot shall correspond to one , and statistical analyses are done on these control-treatment pairs (i.e. demonstrating eligibility for credit issuance by proving statistically significant increase in weathering in treatment plot compared to control plot). The characteristics that shall be measured in control plots are presented in Table 2.

For projects using Method 1: Direct measurement of export to calculate CDR in the NFZ, if the sampling point at the end of the NFZ is catchment or drainage waters, an assessment of the site hydrology shall ensure that catchment or drainage waters of the treatment and control plots remain separate. This is to avoid collecting water that mixes signals from both plots, as this would compromise the comparison.

Table 2 The characteristics to measure in control plots and the measurement frequency.

Characteristic	Frequency	Note
Baseline CDR from counterfactual weathering	Each reporting period	Using the same NFZ measurement method as the treatment area, and FFZ deductions (see GHG quantification section for more details)
Soil organic carbon changes	At least once after rock spreading during the crediting period	Measurement shall be taken during the same reporting period for the treatment area/project scenario
Crop yields	At least once after rock spreading during the crediting period	Measurement shall be taken during the same reporting period for the treatment area/project scenario

Control plot management

Control plots should include business as usual (BAU) practices, including but not limited to continued use of pH management/agricultural lime on agricultural fields at pre-deployment rates. Where this is not possible, negative control plots can be used instead. Negative control plots may include no pH management, but continue other BAU agronomic practices (e.g. cropping, tilling, fertilizer...).

If a negative control plot is used instead of a BAU control plot, it shall be conservatively assumed that all agricultural lime dissolves and generates CDR at 100% efficiency, with negligible carbon loss terms. See the Baseline scenario GHG quantification section for more details.

BAU plots shall maintain the following counterfactual practices where relevant:

• Liming

• Crop selection

• Tilling

• Fertilization

• Irrigation

Project Developers shall justify the amount, frequency, type, and any other relevant information for each BAU practice. The hierarchy of evidence from most to least preferred is:

1. records of historical or recent management/agricultural practices, using a trend/projection of recent practices, unless the Project Developer can justify that the trend is not representative

2. records of historical or recent management/agricultural practices, using a different value within the range of recent practices, e.g. if recent trends are not representative

3. records of average local/regional practices

4. recommended practices by local agronomists or extension agents

Control plot number and size

Each treatment plot shall have a corresponding control plot that is representative of the treatment plot. The number and size of control plots shall be large enough to capture the baseline variability of the corresponding treatment plot, and to ensure statistically significant comparisons with the treatment plot.

At least one control-treatment pair is required per stratum. If control plots are large, it is recommended to split them into multiple smaller plots, rather than maintaining a single, sparsely-sampled large plot. Control plots do not need to be contiguous with one another, or with their corresponding treatment plots. Small projects with only one stratum may contain only one control plot.

Control plots shall represent at least 1% of the total area of a given stratum, with a larger percentage recommended for smaller strata, up to 5%. Based on these guidelines, the following decay function was interpolated (shown in Figure 1) and is recommended to help guide Project Developers in determining the appropriate control plot area as a function of stratum area (in hectares, $x$):

$$\textbf{(Eq. 1)}\ \%\ Control\ area= 4.13 \times e^{-3.14e-4x}$$

Project Developers shall describe in the PDD the number and relative area of control plots in each stratum, considering the size of the stratum, baseline variability, sampling density, and statistical power.$$

Figure 1 The recommended control plot size is plotted against the corresponding stratum size.

Control plot representativeness

Control plots must be representative of the strata to which they belong. A control plot is considered sufficiently representative if the standardized mean difference (SMD) between treatment and control plots across the quantified soil and site characteristics listed in Table 1 is less than 10%.

Alternative justifications for representativeness may be accepted on a case-by-case basis, subject to review. This requirement is expected to be readily met following the stratification steps.

Scaling and extrapolation

The use of and sampling treatment plots already allows Project Developers to scale the size of their deployments with non-linearly increasing sampling requirements. Project Developers may further scale their deployments with one of the following options.

Reduce sampling on existing treatment plots

Conditions: Project Developers are eligible for this option if they meet both of the following requirements for NFZ CDR measurements in the HD sampling plot, after 12 months of monitoring:

1. Low inter-plot variability, demonstrating that strata are sufficiently narrowly defined and that characteristics affecting CDR have been properly considered. This can be shown using tests including, but not limited to Relative Standard Deviation (RSD) / Coefficient of Variation (CV) ≤ 10%.

2. Similarity to the LD plot, demonstrating that the HD plot is representative of the LD plot. This can be shown using tests including, but not limited to t-tests or ANOVA.

Scaling protocol: Reduce sampling density in existing LD plots to a minimum of:

• 1 soil sample per 30 ha (0.033 samples/ha), or

• 1 aqueous sample per 90 ha (0.011 samples/ha),

• or 3 samples per LD treatment plot (whichever is larger) for aqueous and/or soil samples.

Project Developers shall notify Rainbow of their intent to reduce sampling density in LD plots prior to implementation. Eligibility for this adjustment must be approved by a VVB. Project Developers should include this analysis in a Monitoring Report, along with their proposed adjustments to the Sampling Plan, for review and approval by the VVB as part of an ongoing verification audit.

Spread on new treatment plots

Project Developers may use models coupled with empirical measurements to estimate net NFZ CDR on new fields, after validating the model's predictive capacity for a given strata type.

Conditions:

1. Models shall be initialized and/or calibrated using site-specific data for a given stratum.

2. Models shall be spatially and temporally validated using measurements from at least 12 months on the HD and LD treatment plots, i.e.:

   1. Spatial Validation: A model must demonstrate its ability to accurately predict either total CDR across the entire treatment plot (i.e. a general process-based ERW model), or to predict LD plot measurements using HD plot results as input parameters or for calibration (i.e. a data-driven model). This validates the model’s spatial predictive capability, within the same time frame and weather conditions.

   2. Temporal Validation: A model must demonstrate the ability to accurately extrapolate measurements over time, whether from one season to another or from one year to the next for the same season, based on the Project Developer’s objectives. For example, the model should reliably adjust Year 1 HD plot measurements to account for Year 2 weather and agronomic conditions, thereby predicting Year 2 HD results with precision. This validated the model’s temporal predictive capability.

3. The overall approach shall account for variations in feedstock. This can be achieved in one of two ways:

   1. The model is designed to incorporate varying feedstock characteristics as input data, or

   2. The model is initialized and validated for a specific feedstock, and Project Developers limit new spreading events to the use of the same feedstock (i.e., identical particle size distribution, mineralogy, and application rate).

4. Further model requirements are described here.

Scaling protocol: Project Developers may spread feedstock on new fields of the same validated stratum, monitoring them according to the low sampling density presented in the Reducing sampling on existing treatment plots section. Addition of these extrapolated fields to the treatment area does not change the total stratum size, and the required control plot area. The originally established control plot may still be used to compare to the newly established LD treatment plot.

Monitoring results are to be used as a validation check, ensuring that measured CDR values fall within the range of modeled predictions. Any discrepancies must be justified and addressed in the monitoring report. CDR for the entire stratum shall be counted using measurements from the HD and LD treatment plots, and the new extrapolated plot.

Figure 2 The scaling approach using models to extrapolate to new treatment plots is summarized. The model is validated using empirical results in year 1 and 2, for a given strata type with homogeneous soil characteristics. Then in year 3 and beyond, the model may be used to calculate net NFZ removals on newly established treatment plots with low sampling density.