An Integrated CNN Model for Reconstructing and Predicting Land Use/Cover Change: A Case Study of the Baicheng Area, Northeast China

Abstract

Land use and land cover change (LUCC) modeling has continuously been a major research theme in the field of land system science, which interprets the causes and consequences of land use dynamics. In particular, models that can obtain long-term land use data with high precision are of great value in research on global environmental change and climate impact, as land use data are important model input parameters for evaluating the effect of human activity on nature. However, the accuracy of existing reconstruction and prediction models is inadequate. In this context, this study proposes an integrated convolutional neural network (CNN) LUCC reconstruction and prediction model (CLRPM), which meets the demand for fine-scale LUCC reconstruction and prediction. This model applies the deep learning method, which far exceeds the performance of traditional machine learning methods, and uses CNN to extract spatial features and provide greater proximity information. Taking Baicheng city in Northeast China as an example, we verify that CLRPM achieved high-precision annual LUCC reconstruction and prediction, with an overall accuracy rate 9.38% higher than that of the existing models. Additionally, the error rate was reduced by 49.5%. Moreover, this model can perform multilevel LUCC classification category reconstructions and predictions. This study casts light on LUCC models within the high-precision and fine-grained LUCC categories, which will aid LUCC analyses and help decision-makers better understand complex land-use systems and develop better land management strategies.

Predicting LULC changes and assessing their impact on surface runoff with machine learning and remote sensing data.

“…The applied approach used in training the prediction model in python 3 for all algorithms was based on LULCC using historical LULC data and both fixed and temporary spatial factors: Machine (SVM) algorithm even after 120 hours of training. This aligns with the findings of a previous study (Zhang et al 2021), where the authors reported that SVM did not yield results even after more than 200 hours of computation. This suggests that SVM is not suitable for large-scale reconstruction and prediction of LULCC, especially when dealing with extensive datasets.…”

“…Specifically, these studies only update the LULC data in the model input for each prediction while keeping the same temporal factors for a specific year. With the exception of the study by (Zhang et al 2021), which considered the temporality of factors, where the process was followed by temporal selection and con-figuration, resulting in a feature database. In comparison, LULC data utilized by this study from 2000 to 2020, in addition to 11 environmental and socio-economic factors that may impact land use changes shown in Population and rainfall variables were generated using the interpolation method known as Inverse Distance Weighting (IDW), expressed as follows:…”

Land Use/Land Cover Changes in Baicheng District, China during the Period 1954–2020 and Their Driving Forces

“…LULC research has become a global research hotspot [23]. From the analysis of LULC spatial-temporal patterns [24][25][26] and explorations of LULC change driving mechanisms [27,28], the research has evolved to focus on the impact on ecosystem services [29,30], future LULC simulation studies [31][32][33][34], and more detailed analyses of human activities, policies, and other LULC driving mechanisms [35]. In terms of research scale, there have been numerous studies at global, continental, and transnational levels.…”