Rain Patterns Research Articles

Enhancing the Rainfall Forecasting Accuracy of Ensemble Numerical Prediction Systems via Convolutional Neural Networks

Abstract Ensemble prediction systems are commonly used to demonstrate the potential uncertainty of weather forecasts. These systems also help provide weather predictions to prepare for disasters. Specifically, a type of prediction called a quantitative precipitation forecast (QPF) is calculated from the average of multiple forecasts. The probability-matched mean (PMM) method is used to improve these QPF predictions when they underestimate rainfall. However, the PMM method can exhibit limitations when dealing with certain types of rain patterns. To address this issue, this study focuses on improving short-term heavy rainfall predictions using an artificial intelligence (AI) algorithm that combines deep neural networks (DNNs), including convolutional neural networks (CNNs), with a space-based attention mechanism (convolutional block attention module [CBAM]). Four experiments were conducted to evaluate the new method, including those performed to optimize the timing of a 24-hour QPF model, adjust the training dataset, and refine the training algorithm. Two specific weather events were assessed: rainfall influenced by southwest winds after typhoons and afternoon thunderstorm rainfall. A comparison of the results obtained via the root mean square error (RMSE) and structural similarity index measure (SSIM) reveals that the accuracy and distribution of the QPF predictions significantly improved over those of the PMM method. Compared with the PMM method, our approach can reduce the RMSE from 77.51 to 19.52 in the Mei-yu case, an improvement of approximately 74.82%. The SSIM also increased from 0.33 to 0.56, indicating a 70.9% enhancement. Overall, the new approach successfully enhances rainfall prediction accuracy via AI techniques and has the potential to be applied in disaster preparedness operations.

GIS-based method for assessing the viability of solar-powered irrigation

To attain food security, sub-Saharan Africa needs to increase its use of irrigation to improve agricultural productivity and resilience to climate change-induced droughts and erratic rain patterns. Stand-alone solar-powered irrigation systems are a promising technology to power irrigation where grid electrification is unavailable. However, these systems’ economic viability and sustainability are affected by spatially varying factors such as climate, water availability, and solar potential, making it challenging to plan their roll-out. This study addresses the planning challenge by developing a novel open-source reproducible geospatial methodology to determine irrigation water demand, peak power demand, life cycle cost, and locations where solar-powered irrigation systems would be cheaper than alternative off-grid irrigation technologies. The method considers spatial–temporal data for meteorological conditions, soil quality, water distance and depth, solar resource potential, technology efficiency and local equipment and fuel costs. Maize cultivation in Kenya is taken as a case study to demonstrate the method. We find the median peak irrigation water demand in Kenya to be 65,000 l/ha/day and the median peak power demand to be 2.6 kW/ha. The levelized cost of solar-powered irrigation systems ranges from US$ 0.09/kWh to US$ 0.25/kWh, making solar irrigation cheaper than diesel-powered irrigation in all Kenyan regions. The results from the method offer valuable insights to governments, farmers and investors on where water use regulations are required for sustainable water use, where subsidies and innovative financial models are needed for the affordability of solar irrigation and where alternative uses of energy from the solar systems beyond irrigation are possible.

High-Resolution Rainfall Estimation Using Ensemble Learning Techniques and Multisensor Data Integration.

In Indonesia, the monitoring of rainfall requires an estimation system with a high resolution and wide spatial coverage because of the complexities of the rainfall patterns. This study built a rainfall estimation model for Indonesia through the integration of data from various instruments, namely, rain gauges, weather radars, and weather satellites. An ensemble learning technique, specifically, extreme gradient boosting (XGBoost), was applied to overcome the sparse data due to the limited number of rain gauge points, limited weather radar coverage, and imbalanced rain data. The model includes bias correction of the satellite data to increase the estimation accuracy. In addition, the data from several weather radars installed in Indonesia were also combined. This research handled rainfall estimates in various rain patterns in Indonesia, such as seasonal, equatorial, and local patterns, with a high temporal resolution, close to real time. The validation was carried out at six points, namely, Bandar Lampung, Banjarmasin, Pontianak, Deli Serdang, Gorontalo, and Biak. The research results show good estimation accuracy, with respective values of 0.89, 0.91, 0.89, 0.9, 0.92, and 0.9, and root mean square error (RMSE) values of 2.75 mm/h, 2.57 mm/h, 3.08 mm/h, 2.64 mm/h, 1.85 mm/h, and 2.48 mm/h. Our research highlights the potential of this model to accurately capture diverse rainfall patterns in Indonesia at high spatial and temporal scales.

Simulation and Risk Assessment of Flood Disaster at the Entrance to a Rail Transit Station under Extreme Weather Conditions—A Case Study of Wanqingsha Station of Guangzhou Line 18

With the rapid development of urbanization and underground transportation, as well as the frequent occurrence of extreme weather conditions such as extreme rainfall, flooding disasters for rail transit are becoming severe, and need to be urgently clarified in terms of the mechanism causing them. In this study, a comprehensive model for water damage at the entrance to a rail transit station is proposed, emphasizing the entire process of extreme weather–surface ponding–underground intrusion. The model is validated by the inundation process of Line 5 of the Zhengzhou Metro during the “7.20” event and further applied to Wanqingsha Station of Guangzhou Metro Line 18 in China to determine the surrounding water depth, distribution, total water inflow volume, and water damage time under different rainfall intensities, rain patterns and protection scenarios. It was found that when rainfall reaches the level of a 1-in-2000-years event, the surface water begins to invade the internal rail transit system through the rail transit entrances. When facing extreme rainfall akin to the “7.20” event in Zhengzhou, the rail transit system in Wanqingsha Station meets a heightened risk of water damage, resulting in significantly deeper water levels compared to 1-in-5000-year rainfall event in Guangzhou and exceeds the height of the subway entrances. Analysis of the water intrusion process reveals that, as rainfall intensity escalates, the total inflow water volume into the rail transit system increases while escape time diminishes. Moreover, under identical rainfall intensity, pre-type rainfall yields the highest total water inflow, whereas mid-type rainfall exhibits the shortest escape time. Enhancing the protection conditions can markedly attenuate surface water intrusion into the subterranean rail transit system, thereby enhancing the evacuation time for individuals within the system.

Image Deraining Algorithm Based on Multi-Scale Features

In target detection, tracking, and recognition tasks, high-quality images can achieve better results. However, in actual scenarios, the visual effects and data quality of images are greatly reduced due to the influence of environmental factors, which affect subsequent detection, recognition, and other tasks. Therefore, this paper proposes an image rain removal algorithm based on multi-scale features, which can effectively remove rain streaks. First of all, this paper proposes a deraining algorithm that combines spatial information to improve the network’s generalization ability on real images, aiming at the problem of synthetic datasets used by previous deraining algorithms. Then, by proposing a multi-scale rain removal algorithm, it improves the feature extraction capabilities of existing deraining algorithms. Before extracting deep rain features, a preliminary fusion of multi-scale shallow features can be performed, which can show better performance in images of different sizes. In addition, a spatial attention module and channel are introduced. The attention module increases the ability to extract rain information at each scale; the resulting multi-scale feature image rain removal algorithm is called MFD. Finally, the rain removal algorithm is validated on the rain removal dataset, and the proposed method can effectively remove rain patterns, provide strong performance improvement on several datasets in the image rain removal task, and provide high-quality images for subsequent detection and recognition tasks.

Rainfall Estimation Model in Seasonal Zone and Non-Seasonal Zone Regions Using Weather Radar Imagery Based on a Gradient Boosting Algorithm

Indonesia, a country located in the equatorial region with hilly and valley lands surrounded by vast oceans, has complex rainfall patterns that can generally be classified into three types: equatorial, monsoon, and local. Rainfall estimates have only been derived based on local data and characteristics so far, and have not yet been developed based on universal data for all of Indonesia. This study aimed to develop a rainfall estimation model based on weather radar data throughout Indonesia using ensemble machine learning with the gradient boosting algorithm. The proposed rainfall estimation model is universal, can be applied to different rainfall pattern areas, and has a temporal resolution of 10 min. It is based on determining the root mean square error (RMSE) and R-squared (R2) values. Research was conducted in six areas with different rainfall patterns: Bandar Lampung and Banjarmasin with monsoon rain patterns, Pontianak and Deli Serdang with equatorial rain patterns, and the Gorontalo and Biak areas with local rain patterns. The analysis of the proposed model reveals that the best hyperparameters for the learning rate, maximum depth, and number of trees are 0.7, 3, and 50, respectively. The results demonstrate that the estimated rainfall in the six areas was very accurate, with RMSE < 2 mm/h and R2 > 0.7.

Spatiotemporal Variation and Causes of Typical Extreme Precipitation Events in Shandong Province over the Last 50 Years

In this study, based on hourly ERA5 reanalysis data from July to September, from 1971 to 2020, for Shandong Province, we used mathematical statistical analysis, the Mann–Kendall nonparametric statistical test, cluster analysis, and other methods to extract and analyze the spatiotemporal variation characteristics and causes of typical extreme precipitation events. The results indicated the following: (1) The total number and duration of precipitation events show a nonsignificant upward trend, while the average and extreme rainfall intensities show a nonsignificant downward trend. (2) Extreme precipitation events are primarily concentrated in Qingdao, Jinan, Heze, and Binzhou, with fewer events occurring in central Shandong Province. (3) Extreme precipitation events are classified into four types (namely, patterns I, II, III, and IV). Pattern I exhibits two rain peaks, with the primary rain peak occurring after the secondary rain peak. Similarly, pattern II also displays two rain peaks, with equivalent rainfall amounts for both peaks. In contrast, pattern III has multiple, evenly distributed rain peaks. Finally, pattern IV shows a rain peak during the first half of the precipitation event. Pattern I has the highest occurrence probability (46%), while pattern IV has the lowest (7%). (4) The spatial distributions of the different rain patterns are similar, with most being found in the eastern coastal and western regions. (5) Extreme precipitation events result from interactions between large-scale circulation configurations and mesoscale convective systems. The strong blocking situation and significant circulation transport at middle and low latitudes in East Asia, along with strong convergent uplift, abnormally high specific humidity, and high-water-vapor convergence centers, play crucial roles in supporting large-scale circulation systems and triggering mesoscale convective systems.

Local social-ecological context explains seasonal rural-rural migration of the poorest in south-west Bangladesh

Bangladesh is one of the countries most affected by climate change. Internal migration is often presented as a response to environmental degradation. Here, using a people-centred perspective, we explore the complexity of the links between climate-induced change, environmental degradation caused by waterlogging and seasonal rural migration. We used an inductive qualitative approach in social sciences, conducting fourteen semi-directed interviews and six focus group discussions in March-April 2022. We related those results to a rainfall analysis on CHIRPS data for 1981-2021and we represented interactions and feedback between changes and livelihoods in a model. A complex picture of the situation is emerging, showing the interweaving effects of non-climatic and climatic changes, their interplay at different scales, their cumulative effects, the interactions between livelihood types and feedback between social and natural systems. Most of the climate-induced changes gradually become noticeable over the past 25 years. Climate data confirm these changes in recent decades, with July being wetter and January being dryer. Villagers reported waterlogging as the most significant change in their community, pointing to its multiple causes, originating in non-local and local, non-climatic anthropic changes, exacerbated by shrimp farm enclosures and worsened by climate-induced changes such as heavier rains, wetter monsoons and cyclones. Tiger prawn farms, reported as a lucrative and local adaptation to waterlogging and salinisation for the ones who can afford it, worsen the situation for the less wealthy, causing waterlogging and salinisation of the adjacent agricultural lands and buildings, the disappearance of traditional fishing and a reduction of the local job market. In addition, erratic rain patterns, droughts and cyclones affect local production and labour markets. COVID-19 lockdowns, by impacting markets and mobilities, further aggravated the situation. Inequality has increased as the range of adaptations of the less wealthy appears limited in this context of multiple crises.

Urban inundation rapid prediction method based on multi-machine learning algorithm and rain pattern analysis

Urban inundation disasters caused by extreme rainfall events are becoming increasingly severe. However, the numerical inundation model based on physical process has relatively slow computational speed and struggle to meet the demands of current forecasting and early warning systems. Addressing how to integrate machine learning technology into rapid urban inundation prediction, especially when dealing with limited training samples, and selecting the most appropriate machine learning algorithms for different rain patterns, is a crucial and pressing issue. To tackle this problem, this work combines a coupled hydrological-hydrodynamic model known for its high-computational accuracy with efficient machine learning algorithms. Taking different rain patterns into account, a rapid urban inundation prediction method is proposed. This method utilizes data-driven results from the coupled hydrological-hydrodynamic model and constructs multiple machine learning algorithms based on the conclusions drawn from the analysis regarding the most suitable machine learning algorithms for different rain patterns. The results show that for single-peak, double-peak, and uniform rain patterns, the Ridge algorithm, KNN algorithm, and RF algorithm should be respectively employed. The method achieves a mean absolute percentage error of 5.32 %, 7.73 %, and 2.49 % for single-peak, double-peak, and uniform rainfall scenarios and can predict inundation within a 3.68 km2 area in 14.07 s. This method offers high prediction accuracy and fast computational speed, not only meeting the requirements of daily forecasting and early warning but also providing a new direction for the application of machine learning technology in urban inundation prevention.

Mitigation and adaptation of local organization to achieve food security: Case of traditional irrigation system in Bali, Indonesia

Climate change has affected the agricultural sector, particularly the food crop cultivated on rice fields. Subak as a local farmers organization is always expected to have important role to to improve rice farming practices and increase its productivity. The changes of rainfall and rain patterns happened in Bali should be anticipated by the subak through the mitigation and adaptation efforts to prevent the failure of harvest, and ensure better production. The aims of this study were to describe subaks activities and its problem, and the adaptation and mitigation efforts for achieving food security. The study was purposively conducted in 14 subaks which obtained irrigation water from the Buleleng River, Bali Province. A total of 140 farmers were selected as samples by using non-proportional sampling technique. Primary data and secondary data were collected using survey, direct observation, and FGD. Data collected were analyzed by a qualitative descriptive method. The results indicate that subaks have some problems, such as water availability, competition of water among the users, and rice field conversion. The efforts of adaptation and mitigation comprise the changes in cropping patterns and planting schedules, borrowing water between subak members, and forming a subaks federation to increase rice productivity and achieve food security programs. The government should provide useful weather information for farmers regarding water availability, make a digital map of forest, regulations, and law enforcement.

Aedes aegypti oviposition dynamics in towns with low human population density in yungas and dry chaco, Salta, Argentina

We assessed the presence of Aedes aegypti in five ecorregions of Salta province and compared the oviposition activity of Ae. aegypti using ovitraps in towns of two contrasting ecoregions (yungas and Chaco dry forests) in the province of Salta, Argentina, a major contrast in these ecoregions are rain patterns and altitude. Our aim was to estimate how oviposition activities were associated with the ecoregion and site scale local environmental variables. Mosquito oviposition activity was monitored weekly during the summer using ovitraps. Predictor variables were ecoregion, town, and meteorological variables. The effect of the predictor variables was measured on the response variables using multi-model inference. Besides yungas, the presence of Aedes aegypti was confirmed in towns of dry Chaco and High Monte. The only factor that had a significant effect on the presence of eggs in the ovitraps was the ecoregion, with the frequency of positives being higher in yungas. For the number of eggs, the ecoregion, the night temperature of the first week and the NDVI would explain said variable. Overall, results indicate that the variations between towns would be more related with their ecological and climatic characteristics than with the more immediate meteorological variations.

Lasting effects of avian‐frugivore interactions on seed dispersal and seedling establishment

Abstract The consequences of plant–animal interactions often transcend the mere encounter stage, as those encounters are followed by a chain of subsequent stages on the plant's reproductive cycle that ultimately determine fitness. Yet, the dissemination and recruitment stages of animal‐mediated seed dispersal are seldom analysed jointly, hindering a full understanding of the ecology of seed dispersal. We analyse the dispersal and recruitment of a fleshy‐fruited plant (Pistacia lentiscus), from fruit production to seedling survival up to their second year. We link early reproductive investment of individual plants to seedling recruitment and explore the role played by seed viability, the coterie of frugivores and microhabitat seed deposition. The proportion of viable seeds was generally low (mean = 34%) but highly variable among individual plants (range: 0%–95%). Seed viability did not seem to have a direct effect on individual plant's recruitment. We recorded 28 bird species feeding on P. lentiscus fruits or seeds. Their contribution to plant recruitment was mainly determined by their intensity of fruit consumption and probability to disperse viable seeds. Most frugivores presented non‐random microhabitat preferences, delivering uneven seed contributions to different sites. Post‐dispersal seed predation by rodents was the most limiting phase in P. lentiscus recruitment. Yet, microhabitats showing the lowest predation rates received the lowest seed rain. Hence, we found a decoupling of the dissemination and recruitment stages: most seeds do not arrive at the most suitable microhabitats. We estimate P. lentiscus plants need to produce c. 5 × 105 fruits to recruit a single seedling that survives to its second summer in our study site. Its success as a prevalent species in Mediterranean lowland landscapes relies on its high fecundity and thorough fruit removal and dispersal by a diversified frugivore assemblage, which compensates for the high seed unviability characteristic of this genus. Synthesis: Measuring the delayed, post‐dispersal outcomes of animal frugivory interactions may overturn inferences based on consumption observations only. Seed rain patterns are often decoupled from microhabitats' suitability for seedling recruitment. Hence, more integrative studies that encompass the entire plant reproductive cycle (from fruit production to seedling recruitment) are needed to fully understand frugivores' lasting contribution to plant regeneration in natural populations.

Classification of tropical cyclone rain patterns using convolutional autoencoder

Heavy rainfall produced by tropical cyclones (TCs) frequently causes wide-spread damage. TCs have different patterns of rain depending on their development stage, geographical location, and surrounding environmental conditions. However, an objective system for classifying TC rain patterns has not yet been established. This study objectively classifies rain patterns of North Atlantic TCs using a Convolutional Autoencoder (CAE). The CAE is trained with 11,991 images of TC rain rates obtained from satellite precipitation estimates during 2000−2020. The CAE consists of an encoder which compresses the original TC rain image into low-dimensional features and a decoder which reconstructs an image from the compressed features. Then, TC rain images are classified by applying a k-means method to the compressed features from the CAE. We identified six TC rain patterns over the North Atlantic and confirmed that they exhibited unique characteristics in their spatial patterns (e.g., area, asymmetry, dispersion) and geographical locations. Furthermore, the characteristics of rain patterns in each cluster were closely related to storm intensity and surrounding environmental conditions of moisture supply, vertical wind shear, and land interaction. This classification of TC rain patterns and further investigations into their evolution and spatial variability can improve forecasts and help mitigate damage from these systems.

Spatial patterns of torrential rain in the Haihe River Basin and the corresponding large‐scale circulation

AbstractThe classification of torrential rain is crucial to the flood mitigation and water resources, but the relevant research in the Haihe River Basin (HRB) is lacking. In this study, 156 regional torrential rain events from 1972 to 2022 were classified into two types using cluster analysis: eastern type and southern type. The eastern type is the typical torrential rain in the HRB, accounting for 61.5% of the total days, whereas the southern type mainly occurs in the piedmont plain with larger intensity and scope, which would pose more severe hazards. Corresponding large‐scale circulation patterns of the two types reveal a clear relationship between the HRB torrential rain and ascending motion, associated with the anomalies of low‐level moisture convergence and upper‐level wind divergence. The anomalous upper‐level divergence is accompanied by the northward expansion of South Asian High and intensified westerly jets. In the eastern type, strong low‐pressure anomalies north of Lake Baikal at mid‐level produce anomalous northwesterlies. The cold air from mid–high latitudes encounters enhanced southwesterly moist air accompanied by the northward expansion of Western Pacific Subtropical High (WPSH), resulting in the low‐level moisture convergence. In the southern type, the low‐level anomalous convergence is caused by a relatively weak cyclonic anomaly and significantly strengthened southeasterly moisture from the Pacific. The enhancement of southwesterlies results from the cooperative effect of the remarkably northwesterly WPSH and low‐pressure anomalies in South China Sea. As a whole, the eastern type is the result of convergence of cold air and southwesterly moisture driven by a dipole pattern at mid–high latitudes, whereas the southern type is mainly attributed to the accumulation of abundant southeasterly water vapour. This classification can identify high‐risk areas for torrential rain in the HRB and contribute to the forecast skills of extreme precipitation.

A review of natural and managed revegetation responses in two de-watered reservoirs after large dam removals on the Elwha River, Washington, USA

Large dam removals are increasing in frequency and the response of natural and managed revegetation is a critical consideration for managed restoration of dewatered reservoir landscapes post dam removal. The removal of two large dams on the Elwha River in 2011-2014 provides insight into reservoir revegetation. We review literature and datasets from 2012 through 2018, 1-6 years since reservoir dewatering, to compare pre-dam removal predictions on the Elwha to post-dam removal of natural revegetation, managed revegetation effects and invasive non-native vegetation response. Pre-dam removal hypotheses about natural revegetation did not predict species performance on reservoir sediments, seed rain patterns, or seed bank response. Sediment texture and landform affected multiple aspects of revegetation, including vegetation cover, species richness, woody stem densities and species composition. Reservoir drawdown timing influenced species composition and seedling densities. Predictions about managed revegetation effects were mixed. Planting trees and shrubs did not accelerate woody cover but did increase species richness. Seeding reduced non-native vegetation frequency and species richness, had no effect on vegetation cover on fine sediments, but increased vegetation cover on coarse sediments. Planting trees and shrubs during drawdown appeared to result in higher survival rates compared to plantings installed 1+ years post drawdown. Seeding Lupinus rivularis (riverbank lupine) on coarse sediments was successful and increased foliar nitrogen in planted conifers. Invasive non-native vegetation was correctly predicted to be more abundant in the Aldwell reservoir but did not preclude native species establishment in either reservoir, likely due to rapid establishment of native species and robust management that occurred before, during and after dam removal.

Community Empowerment in Drainage Management to Overcome Floods

Flooding is a serious problem that often disrupts the lives of urban communities. Increased rainfall intensity, climate change, and rapid urban development have made existing drainage systems less effective at coping with flooding. To face this challenge, this study aims to examine community empowerment efforts in drainage management as a flood mitigation strategy. This research was conducted by combining participatory approaches and field research methods. Local communities, local governments, and stakeholders are involved in planning, implementing, and evaluating empowerment programs. This program includes skills training, education on the importance of drainage maintenance, as well as the formation of community groups active in drainage maintenance. The results showed that community empowerment in drainage management has had a positive impact in reducing flood risk. Community groups involved in this program actively maintain the cleanliness and drainage function in their neighborhoods. In addition, education about climate change and rain patterns also helps people to be better prepared for extreme conditions.

An Analytical Model for Assessing Dynamic Stability of Bedding Rock Slope with Soil Interlayer Under Different Rain Patterns

An Analytical Model for Assessing Dynamic Stability of Bedding Rock Slope with Soil Interlayer Under Different Rain Patterns

Climate Change Variation And Farmer’s Viewpoint On Agriculture In Semi-Arid Region

It is anticipated that the variation in change within climate brought on by an increase in greenhouse gases would lead to greater temperatures as well as changing patterns of rain and snow, both of which will have a momentous impact on the lives of people. This research was conducted with the objectives of determining how residents of northern Nigeria's farming communities felt about the change within climate variation affects, how they reacted to those changes, and what impact those changes had on agricultural output in the region. For the purpose of the study, one hundred residents of Kaduna were selected at random for taking part. In order to learn how the unpredictability of change within climate and the strategies of adaptation impact agricultural production, a descriptive poll was conducted. The findings indicate that greater majority of individuals are aware of the gravity of change within climate and the fact that the majority of the impacts of change within climate on agricultural and livestock output are negative. The majority of individuals believe that the many strategies that have been developed to counter the undesirable effects of change within climate on agricultural output and livestock production are effective, as the findings of the survey demonstrated.

Analysis of the Relationship between Extreme Rain Patterns and Maximum Runoff in Xuyi County Urban Area: Production Confluence of Urban Underlying Surface

The relationship between extreme rainfall and maximum runoff in Xuyi County, Huai’an City, Jiangsu Province, China was studied by using the Chicago rain pattern method, statistical analysis, and literature review. From 1957 to 2020, the average precipitation was 1015 mm with a fluctuating upward trend mainly in July-August which accounted for 40% of the total annual precipitation. The maximum evaporation occurred in June, accounting for 12.7% of the total annual evaporation. From 1957 to 2020, the average maximum daily precipitation was 45 mm, while the maximum precipitation was found as 233.2 mm on July 27, 1970. In the 120-minute rain pattern, the maximum precipitation coefficient was 0.4, and the rainfall intensity showed an unimodal distribution. The total annual runoff control rate increased from 60 to 95%, the corresponding design precipitation increased from 16.3 to 71.5 mm, and the growth rate of design precipitation became faster.

Background-detail restoration image deraining network based on convolutional dictionary network

The heavy rain streaks can seriously fade the color of images and destroy the background texture information. Although the existing deraining methods perform well in rain removal, we find that most of them tend to favor rain removal or restoration. In this paper, we propose a background-detail restoration image deraining network based on convolutional dictionary network to achieve effective deraining results with better-restored background details. Our framework includes two sub-networks, namely the rain-removal network and the background-restoration network. The former sub-network is constructed as the multi-stage architecture. In order to accurately maintain as much background detail as possible while removing rain streaks, we break down the overall rain removal process into more manageable steps. We introduce convolutional dictionary network and Multi-Scale Detail Restoration Block (MSDRB) to achieve the balance between the effectiveness of deraining and background restoration. The second sub-network is Residual-Unet with Squeeze-and-Excitation (RUSE) operation, which can utilize the output of every stage generated from the former sub-network to fine-tune the background details in rain-free images. Experiments on benchmark public synthetic datasets and real-world datasets demonstrate the effectiveness and generalization capability of the proposed method in rain streak removal and background restoration with different rain patterns.