HYDROPEDOLOGICAL PROCESSES AND THEIR IMPLICATIONS FOR PRECISION AGRICULTURE
Open Access
- Author:
- Zhu, Qing
- Graduate Program:
- Soil Science
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- September 21, 2009
- Committee Members:
- Hangsheng Lin, Dissertation Advisor/Co-Advisor
Hangsheng Lin, Committee Chair/Co-Chair
Jason Philip Kaye, Committee Member
John Earl Watson, Committee Member
John Matthew Schmidt, Committee Member
Kamini Singha, Committee Member - Keywords:
- Pedology
Hydrology
Geostatistics
Soil property
Soil moisture
Soil Physics - Abstract:
- Within-field spatial heterogeneity of soil and water are widely spread in the real world. However, means for detecting and mapping the spatial heterogeneity are still limited. Effects of the spatial heterogeneity on water and fertilizer transport, transformation, and storage also have not been fully understood. The objectives of this study are two fold: 1) to understand hydropedological functions in an agricultural landscape typical of central Pennsylvania, including mapping and identifying soil moisture variation, subsurface flow paths, and soil properties spatial patterns, and 2) to examin how these quantified hydropedological functions affect soil nitrogen (N) availability to corn in this agricultural landscape, with implications for precision agriculture. First, we investigated appropriate kriging methods for mapping different point-based hydropedological properties over the landscape. We examined the combined impacts of sample size, spatial structure, and auxiliary variables on the interpolation accuracy for diverse soil properties and landscape features. The results showed that regression kriging (RK) is preferable when spatial structure can not be well captured (e.g., the ratio of sampling spacing over range > 0.5) or a strong relationship exists between target properties and auxiliary variables (e.g., R2 > 0.6); otherwise, ordinary kriging (OK) performs better for the hydropedological properties investigated in this landscape with multiple parent materials and gentler slope. Soil water movement and its controlling factors were investigated to understand hydropedological processes in relation to soil-landscape features. The deterministic 8 single-flow (D8) algorithm was used to simulate lateral concentrated subsurface flow paths at the interfaces of Ap1–Ap2 soil horizon, clay layer (>40% clay cotent), and soil–bedrock. The results confirmed that subsurface concentrated lateral flow only occurred at the interfaces of the clay layer and the soil–bedrock. At these two interfaces, soils on the simulated flow paths were closer to saturation. Higher Mn content was observed on the simulated flow paths at the two interfaces. Soils on the simulated flow paths also had higher soil apparent electrical conductivity readings. We then investigated the factors influencing soil moisture variations at four spatial scales in this landscape, including the entire farm (19.5 ha), landform unit (0.3-5.1 ha), plot size (10×10 m), and hillslope transect (40×2 m). The effects of soil properties, terrain attributes, and crop growth at multiple depths and in different seasons (growing vs. non-growing) were examined. The results showed that crop and soil influences on soil moisture dynamics were stronger during the growing seasons compared to the non-growing seasons, while topography influences showed an opposite effect. The topographic influences on soil moisture became more dominant as topography contrast or soil depth increased in the study area. Next, we conducted repeated electromagnetic induction (EMI) surveys using different meters (EM38, EM31, and Dualem-2), dipole orientations (V and H), and geometries (HCP and perpendicular-PRP) to explore the best possible protocol for detecting soil and hydrological variations over the landscape. The results suggested that mapping different hydropedological properties using EMI should consider optimal timing (e.g., during wetter period), meter, dipole orientation, and geometry. Repeated EMI surveys within a few months revealed soil moisture temporal dynamics. Correlation between apparent electrical conductivity (ECa) and soil moisture was higher (R2 = 0.59-0.77) in wetter areas. The EM31V with 6-m measurement depth had the best correlation to depth to bedrock (R2 = 0.58). Because the top 2 m of most soil profiles exhibited distinct differences in texture, the EM31H and EMHCP with 3-m measurement depth showed the best correlation with soil silt content (R2 > 0.45). Finally, this study examined the relationships between corn response to N fertilizer and related hydropedological processes. The experiment was conducted at three sites (labeled as A, B, and C), which represent distinct landforms (i.e., depressional area, steep slope, and flood plain), soil types (i.e., Hagerstown, Opequon, and Melvin series), and hydrological features (including drainage, subsurface flow, and Ksat). A replicated small-plot study was conducted in each of the chosen three sites, including four blocks of two N treatments (NH4NO3) applied to corn at planting (0 and 150 kg N ha-1). The results showed that Site A had the greatest grain yield and biomass because of its fine texture (>19% clay) and low Ksat (<0.3 cm min-1). Site B had the least grain yield, corresponding with the least inorganic soil N content because of its thin Ap1 and Ap2 horizons and shallow lateral subsurface flow at the soil-bedrock interface (0.4 m below the surface). Despite the higher amount of inorganic soil N content at Site C, the biomass and grain yield were small and resulted from the drier soil during the growing season.