Month: May 2024

Abstract

India experienced several unprecedented floods in the recent decades. The increase in the extreme rainfall events (EREs) is the primary cause for these floods, manifesting its societal impacts. The daily downscaled and bias corrected (DBC) Coupled Model Intercomparison Project Phase 6 (CMIP6) rainfall and sea surface temperature (SST) are prepared for the Indian region and are utilized to examine the characteristics of EREs. The DBC products capture the characteristic features of EREs for the baseline period, which inspired us to assess the EREs over India in CMIP6 future projections. Consistent with the observations, DBC product shows ~ 8% of Indian land found to experienced extremely heavy rainfall associated with the long duration EREs in the baseline period. However, area and extreme rainfall thresholds are projected to increase by about 18(13)% and 58(50)%, respectively in the far future under SSP5-8.5 (SSP2-4.5) emission scenario relative to the baseline period. A two-fold-65(62)% increase in long-duration EREs compared to the short-duration EREs and substantial warming ~ 2.4(2.9) ^oC of Indian Ocean SSTs in the far future under SSP5-8.5 (SSP2-4.5) emission scenario compared to baseline period are reported. These findings may provide fundamental insights to formulate national climate change adaptation policies for the EREs.

Abstract

An accurate understanding of the current and future water cycle over the Third Pole is of great societal importance, given the role this region plays as a water tower for densely populated areas downstream. An emerging and promising approach for skillful climate assessments over regions of complex terrain is kilometer-scale climate modeling. As a foundational step towards such simulations over the Third Pole, we present a multi-model and multi-physics ensemble of kilometer-scale regional simulations for the hydrological year of October 2019 to September 2020. The ensemble consists of 13 simulations performed by an international consortium of 10 research groups, configured with a horizontal grid spacing ranging from 2.2 to 4 km covering all of the Third Pole region. These simulations are driven by ERA5 and are part of a Coordinated Regional Climate Downscaling EXperiment Flagship Pilot Study on Convection-Permitting Third Pole. The simulations are compared against available gridded and in-situ observations and remote-sensing data, to assess the performance and spread of the model ensemble compared to the driving reanalysis during the cold and warm seasons. Although ensemble evaluation is hindered by large differences between the gridded precipitation datasets used as a reference over this region, we show that the ensemble improves on many warm-season precipitation metrics compared with ERA5, including most wet-day and hour statistics, and also adds value in the representation of wet spells in both seasons. As such, the ensemble will provide an invaluable resource for future improvements in the process understanding of the hydroclimate of this remote but important region.

Abstract

Existing image super-resolution methods have made remarkable advancements in enhancing the visual quality of real-world images. However, when it comes to restoring Chinese paintings, these methods encounter unique challenges. This is primarily due to the difficulty in preserving intricate non-realistic details and capturing comple semantic information with high dimensionality. Moreover, the preservation of the original artwork’s distinct style and subtle artistic nuances further amplifies this complexity. To address these challenges and effectively restore traditional Chinese paintings, we propose a Convolutional Super-Resolution Generative Adversarial Network for Chinese landscape painting super-resolution, termed ConvSRGAN. We employ Enhanced Adaptive Residual Module to delve deeply into multi-scale feature extraction in images, incorporating an Enhanced High-Frequency Retention Module that leverages an Adaptive Deep Convolution Block to capture fine-grained high-frequency details across multiple levels. By combining the Multi-Scale Structural Similarity loss with conventional losses, our ConvSRGAN ensures that the model produces outputs with improved fidelity to the original image’s texture and structure. Experimental validation demonstrates significant qualitative and quantitative results when processing traditional paintings and murals datasets, particularly excelling in high-definition reconstruction tasks for landscape paintings. The reconstruction effect showcases enhanced visual fidelity and liveliness, thus affirming the effectiveness and applicability of our approach in cultural heritage preservation and restoration.

Abstract

Assessment of current and future growth in the global rooftop area is important for understanding and planning for a robust and sustainable decentralised energy system. These estimates are also important for urban planning studies and designing sustainable cities thereby forwarding the ethos of the Sustainable Development Goals 7 (clean energy), 11 (sustainable cities), 13 (climate action) and 15 (life on land). Here, we develop a machine learning framework that trains on big data containing ~700 million open-source building footprints, global land cover, road, and population datasets to generate globally harmonised estimates of growth in rooftop area for five different future growth narratives covered by Shared Socioeconomic Pathways. The dataset provides estimates for ~3.5 million fishnet tiles of 1/8 degree spatial resolution with data on gross rooftop area for five growth narratives covering years 2020–2050 in decadal time steps. This single harmonised global dataset can be used for climate change, energy transition, biodiversity, urban planning, and disaster risk management studies covering continental to conurbation geospatial levels.

Abstract

Data mining and analysis are critical for preventing or mitigating natural hazards. However, data availability in natural hazard analysis is experiencing unprecedented challenges due to economic, technical, and environmental constraints. Recently, generative deep learning has become an increasingly attractive solution to these challenges, which can augment, impute, or synthesize data based on these learned complex, high-dimensional probability distributions of data. Over the last several years, much research has demonstrated the remarkable capabilities of generative deep learning for addressing data-related problems in natural hazards analysis. Data processed by deep generative models can be utilized to describe the evolution or occurrence of natural hazards and contribute to subsequent natural hazard modeling. Here we present a comprehensive review concerning generative deep learning for data generation in natural hazard analysis. (1) We summarized the limitations associated with data availability in natural hazards analysis and identified the fundamental motivations for employing generative deep learning as a critical response to these challenges. (2) We discuss several deep generative models that have been applied to overcome the problems caused by limited data availability in natural hazards analysis. (3) We analyze advances in utilizing generative deep learning for data generation in natural hazard analysis. (4) We discuss challenges associated with leveraging generative deep learning in natural hazard analysis. (5) We explore further opportunities for leveraging generative deep learning in natural hazard analysis. This comprehensive review provides a detailed roadmap for scholars interested in applying generative models for data generation in natural hazard analysis.

Abstract

Localised fruit thinning strategies must be investigated to improve precision crop load management in narrow-canopy, multileader apple trees. This study aimed to determine the effects of within-leader and tree total crop load on leaders’ and trunk’s growth, fruit set, yield, and fruit quality in ‘Ruby Matilda’ apples (marketed as Pink Lady®) over three years. Different crop loads were imposed on two leaders (primary and secondary) of bi-axis trees. Leader and trunk relative growth rate, return fruit set, yield, and fruit quality parameters at harvest were measured. High within-leader crop loads led to a significant increase in yield and reductions in trunk growth, return fruit set, and deterioration of fruit quality parameters except for flesh firmness and starch index. Similar trends were observed in whole-tree relationships. High crop load in secondary leaders had moderate negative effects on trunk growth, yield, and fruit mass of primary leaders; it only marginally affected their return fruit set and had no significant effect (p > 0.05) on their fruit quality. A crop load of 6.8 fruit no. cm⁻² of leader cross-sectional area was estimated to achieve a relatively consistent return fruit set within the same leader. At a whole-tree level, a similar crop load (6.9 fruit no. cm⁻² of trunk cross-sectional area) produced a consistent return fruit set despite its higher variability. These crop loads produced high yields (120 and 111 t ha⁻¹, respectively) and good quality fruit. Using individual leaders as management units is recommended to simplify operations and reduce variability.

Abstract

The global climate models (GCMs) are sophisticated tools for determining how the climate system will respond. However, the output of GCMs has a coarse resolution, which is unsuitable for basin-level modelling. Global climate models need to be downscaled at a local/basin scale to determine the impacts of climate change on hydrological responses. The present study attempted to evaluate how effectively various large-scale predictors could reproduce local-scale rain in 35 different locations in India using artificial neural networks (ANN), change-factors (CF), K-nearest neighbour (KNN), and multiple linear regression (MLR). The selection of predictors is made based on the correlation value. As potential predictors, air temperature, geo-potential height, wind velocity component, and relative humidity at specific mean sea-level pressure are selected. The comparison of four different downscaling methods concerning the reproduction of various statistics such as mean, standard deviation at chosen locations, quantile–quantile plots, cumulative distribution function, and kernel density estimation of the PDFs of daily rainfall for selected stations is examined. The CF approach outperforms the other methods at almost all sites (R² = 0.92–0.99, RMSE = 1.37–28.88 mm, and NSE = –16.55–0.99). This also closely resembles the probability distribution pattern of IMD data.