Post-hoc Evaluation of Sample Size in a Regional Digital Soil Mapping Project
Bibliographic record
Abstract
The transition from conventional soil mapping (CSM) to digital soil mapping (DSM) not only affects the final map products, but it also affects the concepts of scale, resolution, and sampling intensity. This is critical because in the CSM approach, sampling intensity is intricately linked to the desired scale of soil map publication, which provided standardization of sampling. This is not the case for DSM where sample size varies widely by project, and sampling design studies have largely focused on where to sample without due consideration for sample size. Using a regional soil survey dataset with 1791 sampled and described soil profiles, we first extracted an external validation dataset using the conditioned Latin hypercube sampling (cLHS) algorithm and then created repeated (n = 10) sample plans of increasing size from the remaining calibration sites using the cLHS, feature space coverage sampling (FSCS), and simple random sampling (SRS). We then trained random forest (RF) models for four soil properties: pH, CEC, clay content, and SOC at five different depths. We identified the effective sample size based on the model learning curves and compared it to the optimal sample size determined from the Jensen–Shannon divergence (DJS) applied to the environmental covariates. Maps were then generated from models that used all the calibration points (reference maps) and from models that used the optimal sample size (optimal maps) for comparison. Our findings revealed that the optimal sample sizes based on the DJS analysis were closely aligned with the effective sample sizes from the model learning curves (815 for cLHS, 832 for FSCS, and 847 for SRS). Furthermore, the comparison of the optimal maps to the reference maps showed little difference in the global statistics (concordance correlation coefficient and root mean square error) and spatial trends of the data, confirming that the optimal sample size was sufficient for creating predictions of similar accuracy to the full calibration dataset. Finally, we conclude that the Ottawa soil survey project could have saved between CAD 330,500 and CAD 374,000 (CAD = Canadian dollars) if the determination of optimal sample size tools presented herein existed during the project planning phase. This clearly illustrates the need for additional research in determining an optimal sample size for DSM and demonstrates that operationalization of DSM in public institutions requires a sound scientific basis for determining sample size.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.119 | 0.268 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".