TY - JOUR

T1 - Predictive mapping of glacial sediment properties (Bellingshausen Dome, King George Island, Antarctica)

AU - Сулейманов, Азамат Русланович

AU - Низамутдинов, Тимур Ильгизович

AU - Мавлюдов, Булат Рафаэлефич

AU - Абакумов, Евгений Васильевич

PY - 2024/2/1

Y1 - 2024/2/1

N2 - Research in Antarctica is of great importance for understanding the Earth’s climate system and the processes that drive it. This study evaluated the spatial distribution of glacial sediment properties on the Bellingshausen Dome or Collins Ice Cap (King George Island, Antarctica). The particle-size distribution, pH H2O, total organic carbon (TOC), mobile forms of ammonium nitrogen (N–NH4), potassium (K2O), and phosphorus (P2O5) were measured and then spatially modelled using regression kriging (RK) and ordinary kriging (OK) approaches. The terrain attributes (elevation, aspect, slope) derived from a digital elevation model with 10-m spatial resolution and distance from a coast were used as explanatory variables. Multiple linear regression models were fitted to describe the relationships between the covariates and properties. The performance of the models was evaluated by the mean absolute error (MAE), root mean square error (RMSE), coefficient of determination (R2), and Nash–Sutcliffe efficiency coefficient (NSE) indices. Overall, model performance statistics showed that RK models performed better than OK and the spatial patterns of some properties were closely related to the character of the glacier topography. Thus, the application of the RK method in combination with auxiliary environmental covariates improved the accuracy of spatial prediction.

AB - Research in Antarctica is of great importance for understanding the Earth’s climate system and the processes that drive it. This study evaluated the spatial distribution of glacial sediment properties on the Bellingshausen Dome or Collins Ice Cap (King George Island, Antarctica). The particle-size distribution, pH H2O, total organic carbon (TOC), mobile forms of ammonium nitrogen (N–NH4), potassium (K2O), and phosphorus (P2O5) were measured and then spatially modelled using regression kriging (RK) and ordinary kriging (OK) approaches. The terrain attributes (elevation, aspect, slope) derived from a digital elevation model with 10-m spatial resolution and distance from a coast were used as explanatory variables. Multiple linear regression models were fitted to describe the relationships between the covariates and properties. The performance of the models was evaluated by the mean absolute error (MAE), root mean square error (RMSE), coefficient of determination (R2), and Nash–Sutcliffe efficiency coefficient (NSE) indices. Overall, model performance statistics showed that RK models performed better than OK and the spatial patterns of some properties were closely related to the character of the glacier topography. Thus, the application of the RK method in combination with auxiliary environmental covariates improved the accuracy of spatial prediction.

KW - Antarctica

KW - Digital mapping

KW - Geostatistics

KW - Glacier

KW - Kriging

KW - Sediments

UR - https://www.mendeley.com/catalogue/9577dad1-7c7d-3628-a425-9e1fa9350bb5/

U2 - https://link.springer.com/article/10.1007/s12665-024-11440-5

DO - https://link.springer.com/article/10.1007/s12665-024-11440-5

M3 - Article

VL - 83

JO - Environmental Earth Sciences

JF - Environmental Earth Sciences

SN - 1866-6280

IS - 4

M1 - 134

ER -