Effects of moss biocrusts on near-surface soil moisture are underestimated in drylands: Insights from a heat-pulse soil moisture sensor

Abstract

Understanding the physical, chemical, and biological processes and estimating water balance in drylands rely on accurate soil moisture (θ) measurements. Surface soil moisture is influenced by the presence and types of biocrusts because they strongly modify surface soil properties and hydrological processes. Nevertheless, it is still a challenge to precisely measure the θ of such thin biocrust layer and to evaluate biocrusts’ roles in soil water balance. Due to the large sensing range (usually ≥ 5 cm), the most dielectric soil moisture sensors are incapable of accurately measuring near-surface (≤5 cm) θ, although they have been widely used for this purpose. Conversely, the heat-pulse (HP) sensor possibly performs better in measuring surface θ of biocrust layer because of its smaller sensing range (1.2 cm). In this study, we assessed the performance of HP sensor in investigating the surface θ differences between moss-dominated biocrusts and bare soil on the northern Chinese Loess Plateau. We measured surface θ of biocrusts and bare soil for one year at 1 cm depth using HP sensors, and at 2 or 5 cm depths using three common dielectric sensors based on alternative measurement principles. Our results showed that the relationship of θ and heat capacity from the HP sensor was quite well-fitted by a linear regression function (R2 = 0.96), and the θ measured by HP sensor in the field agreed well with the oven-drying θ (root-mean-square error ≤ 0.015 cm3 cm−3 and mean deviation ≤ 0.007 cm3 cm−3). The HP sensor measured a 44% greater θ under biocrusts compared to bare soil throughout the experimental period, in contrast to lesser differences detected with the other sensors (17%–33%). Furthermore, the biocrust increased θ by 100% in the dry season (Nov.−Apr.) and by 40% in the wet season (May−Oct.) in contrast to bare soil based on the HP sensor. Moreover, the θ of biocrusts measured by HP sensor was 0.04 cm3 cm−3 higher than that of bare soil under the periods without rainfall, but the biocrusts increased θ by 0.05–0.10 cm3 cm−3 in contrast to bare soil after rainfall. The apparent differences among different types of sensors were reasonably attributed to their variable sensing ranges and volumes of measured soil. Although the installation depths were different among HP sensor and dielectric sensors, HP sensor was undoubtedly more suitable in measuring near-surface θ and sensitive enough to detect even non-rainfall water inputs (dew, fog, and water vapor sorption) into the soil. In conclusion, the θ differences between moss-dominated biocrusts and bare soil are mostly underestimated by common dielectric sensors compared to HP sensors. Therefore, biocrusts play a more vital role in surface water balance of drylands than we generally recognized before.