Browsing by Autor "Guanhua Zhou"

Now showing 1 - 2 of 2

Deep Learning and Transformer Models for Groundwater Level Prediction in the Marvdasht Plain: Protecting UNESCO Heritage Sites—Persepolis and Naqsh-e Rustam
(Multidisciplinary Digital Publishing Institute, 2025) Peyman Heidarian; Franz Pablo Antezana López; Yumin Tan; Somayeh Fathtabar Firozjaee; Tahmouras Yousefi; H. Salehi; Ava Osman Pour; Maria Elena Oscori Marca; Guanhua Zhou; Ali Azhdari
Groundwater level monitoring is crucial for assessing hydrological responses to climate change and human activities, which pose significant threats to the sustainability of semi-arid aquifers and the cultural heritage they sustain. This study presents an integrated remote sensing and transformer-based deep learning framework that combines diverse geospatial datasets to predict spatiotemporal variations across the plain near the Persepolis and Naqsh-e Rustam archaeological complexes—UNESCO World Heritage Sites situated at the plain’s edge. We assemble 432 synthetic aperture radar (SAR) scenes (2015–2022) and derive vertical ground motion rates greater than −180 mm yr−1, which are co-localized with multisource geoinformation, including hydrometeorological indices, biophysical parameters, and terrain attributes, to train transformer models with traditional deep learning methods. A sparse probabilistic transformer (ConvTransformer) trained on 95 gridded variables achieves an out-of-sample R2 = 0.83 and RMSE = 6.15 m, outperforming bidirectional deep learning models by >40%. Scenario analysis indicates that, in the absence of intervention, subsidence may exceed 200 mm per year within a decade, threatening irreplaceable Achaemenid stone reliefs. Our results indicate that attention-based networks, when coupled to synergistic geodetic constraints, enable early-warning quantification of groundwater stress over heritage sites and provide a scalable template for sustainable aquifer governance worldwide.
Lithium quantification based on random forest with multi-source geoinformation in Coipasa salt flats, Bolivia
(Elsevier BV, 2023) Franz Pablo Antezana López; Guanhua Zhou; Lizandra Paye Vargas; Guifei Jing; Maria Elena Oscori Marca; Maritza Villalobos Quispe; Estefany Antonio Ticona; Neyza Maribel Mollericona Tonconi; Elizabeth Orozco Apaza
• PCA on the geoinformation in the RF model improved the establishment of lithium elements. • MLAs are useful model-based tools for GIS-based lithium mapping. • Soil properties are useful to use as a covariable in lithium prediction models. • Atomic Absorption Spectroscopy results can be a variable in mineral prediction. • Satellite images cannot contain water in salt flats areas for lithium mapping. With the widespread use of Lithium (Li) batteries, there is an urgent demand to explore Li deposits. This study aims to apply efficient remote sensing methodologies with multi-resolution images of Sentinel-2, ASTER, and JILIN GP and soil geoinformation to locate Li by mapping its alterations and directly identifying the minerals that contain it. The principal component analysis (PCA) is applied to the satellite bands to gain an initial understanding and analysis of the temporal and spatial behaviour in the Coipasa salt flats in western Bolivia. Then, we employed the Random Forest (RF) method for defining repeatable mineral exploration targets and predictive modelling of mineral prospecting using geographical information system (GIS) tools. Predictive maps are created that contemplate the source, sediment transport, and chemical deposition processes crucial to produce this mineral. The predictive models used satellite bands and physical and chemical soil grid maps; the training data is extracted from sixteen soil and salt samples evaluated using the results of a laboratory test utilizing Atomic Absorption Spectrophotometry. The results demonstrate the high reliability of our predictive models: the values of the Area Under the Curve (AUC) of the Receiver Operating Characteristic Curve (ROC) are between 0.90 and 0.92. The results also show that the combination of PCA on satellite images and terrain information, such as physical, chemical, and morphological properties of the terrain, improves the model to predict the formation of Li ore within the Coipasa salt flats.