Flood risk is a significant challenge for sustainable spatial planning, particularly concerning climate change and urbanization. Phrasing suitable land planning strategies requires assessing future flood risk and predicting the impact of urban sprawl. This study aims to develop an innovative approach combining land use change and hydraulic models to explore future urban flood risk, aiming to reduce it under different vulnerability and exposure scenarios. SPOT-3 and Sentinel-2 images were processed and classified to create land cover maps for 1995 and 2019, and these were used to predict the 2040 land cover using the Land Change Modeler Module of Terrset. Flood risk was computed by combining hazard, exposure, and vulnerability using hydrodynamic modeling and the Analytic Hierarchy Process method. We have compared flood risk in 1995, 2019, and 2040. Although flood risk increases with urbanization, population density, and the number of hospitals in the flood plain, especially in the coastal region, the area exposed to high and very high risks decreases due to a reduction in poverty rate. This study can provide a theoretical framework supporting climate change related to risk assessment in other metropolitan regions. Methodologically, it underlines the importance of using satellite imagery and the continuity of data in the planning-related decision-making process.

In regular convolutional neural networks (CNN), fully-connected layers act as classifiers to estimate the probabilities for each instance in classification tasks. The accuracy of CNNs can be improved by replacing fully connected layers with gradient boosting algorithms. In this regard, this study investigates three robust classifiers, namely XGBoost, LightGBM, and Catboost, in combination with a CNN for a land cover study in Hanoi, Vietnam. The experiments were implemented using SPOT7 imagery through (1) image segmentation and extraction of features, including spectral information and spatial metrics, (2) normalization of attribute values and generation of graphs, and (3) using graphs as the input dataset to the investigated models for classifying six land cover classes, namely House, Bare land, Vegetation, Water, Impervious Surface, and Shadow. The results show that CNN-based XGBoost (Overall accuracy = 0.8905), LightGBM (0.8956), and CatBoost (0.8956) outperform the other methods used for comparison. It can be seen that the combination of object-based image analysis and CNN-based gradient boosting algorithms significantly improves classification accuracies and can be considered as alternative methods for land cover analysis.