The increasing global demand for sustainable agriculture requires accurate and efficient soil analysis methods. Conventional laboratory techniques are often time‐consuming, costly and environmentally damaging. To address this challenge, we developed and validated locally calibrated mid‐infrared (MIR) spectroscopy models for predicting key soil properties pH, phosphorus (P) and exchangeable cations in soil samples from South Africa's Western Highveld region, using a dataset of 979 soil samples and machine learning algorithms Cubist, partial least squares regression (PLSR) and random forest (RF). A subset of spectra was also submitted to the newly developed Open Soil Spectral Library's (OSSL) prediction models to determine whether global prediction models could be used for local soil property prediction. Accurate predictions for pH, calcium (Ca) and magnesium (Mg), with coefficient of determination (R2) values exceeding 0.76 were obtained with the local calibration algorithms. The predictions for P, potassium (K) and sodium (Na) did not meet the requirements for reliability. Soil spectroscopic prediction models calibrated with local soils outperformed the corresponding global prediction models considered. The OSSL prediction results were inaccurate, with a RPIQ <1, and consistently underpredicted all soil properties. Furthermore, the OSSL collection of prediction models does not include a pH (KCl) model, the routinely used pH measurement method in South Africa. These findings highlight the importance of local calibration for accurate soil property prediction and underscore the need for regional representation in global spectral libraries. This research serves as the first local calibration of MIR spectroscopy models for the Western Highveld region of South Africa and provides a foundation for future local soil property inference model development. It also serves as a potential starting point for a comprehensive South African soil spectral library that can be contributed to global spectral libraries.

中文翻译：

---

南非西部高地地区的土壤光谱模型开发：为什么需要本地数据？


全球对可持续农业的需求不断增长，这需要准确高效的土壤分析方法。传统的实验室技术通常耗时、昂贵且对环境有害。为了应对这一挑战，我们开发并验证了本地校准的中红外 （MIR） 光谱模型，用于预测南非西部高地地区土壤样品中的关键土壤特性 pH 值、磷 （P） 和可交换阳离子，使用包含 979 个土壤样本的数据集和机器学习算法立体主义、偏最小二乘回归 （PLSR） 和随机森林 （RF）。还将光谱子集提交给新开发的开放土壤光谱库 （OSSL） 预测模型，以确定全局预测模型是否可以用于局部土壤特性预测。使用局部校准算法获得 pH 值、钙 （Ca） 和镁 （Mg） 的准确预测，决定系数 （R2） 值超过 0.76。P、钾 （K） 和钠 （Na） 的预测不符合信度要求。用当地土壤校准的土壤光谱预测模型优于所考虑的相应全球预测模型。OSSL 预测结果不准确，RPIQ <1，并且始终低估了所有土壤特性。此外，OSSL 预测模型集合不包括 pH （KCl） 模型，这是南非常规使用的 pH 测量方法。这些发现强调了局部校准对于准确土壤特性预测的重要性，并强调了在全球光谱库中进行区域表示的必要性。 这项研究是南非西部高地地区 MIR 光谱模型的首次局部校准，并为未来的局部土壤特性推断模型开发提供了基础。它还可以作为综合南非土壤光谱库的潜在起点，该库可以贡献给全球光谱库。