Analysis of the Spatiotemporal Patterns of Water Conservation in the Yangtze River Ecological Barrier Zone Based on the InVEST Model and SWAT-BiLSTM Model Using Fractal Theory: A Case Study of the Minjiang River Basin

Fractal and Fractional, Journal Year: 2025, Volume and Issue: 9(2), P. 116 - 116

Published: Feb. 13, 2025

The Yangtze River Basin serves as a vital ecological barrier in China, with its water conservation function playing critical role maintaining regional balance and resource security. This study takes the Minjiang (MRB) case study, employing fractal theory combination InVEST model SWAT-BiLSTM to conduct an in-depth analysis of spatiotemporal patterns conservation. research aims uncover relationship between dynamics watershed capacity ecosystem service functions, providing scientific basis for protection management. Firstly, is introduced quantify complexity spatial heterogeneity natural factors such terrain, vegetation, precipitation Basin. Using model, evaluates functions area, identifying key zones their variations. Additionally, employed simulate hydrological processes basin, particularly impact nonlinear meteorological variables on responses, aiming enhance accuracy reliability predictions. At annual scale, it achieved NSE R2 values 0.85 during calibration 0.90 validation. seasonal these increased 0.91 0.93, at monthly reached 0.94 0.93. showed low errors (RMSE, RSR, RB). findings indicate significant differences Basin, upper middle mountainous regions serving primary areas, whereas downstream plains exhibit relatively lower capacity. Precipitation, terrain slope, vegetation cover are identified main affecting changes having notable regulatory effect Fractal dimension reveals distinct structure which partially explains geographical distribution characteristics functions. Furthermore, simulation results based show increasingly climate change human activities frequent occurrence extreme events, particular, disrupts posing greater challenges Model validation demonstrates that SWAT integrated BiLSTM achieves high capturing complex processes, thereby better supporting decision-makers formulating management strategies.

Language: Английский

---

Establishment and Evaluation of Atmospheric Water Vapor Inversion Model Without Meteorological Parameters Based on Machine Learning

Sensors, Journal Year: 2025, Volume and Issue: 25(2), P. 420 - 420

Published: Jan. 12, 2025

Precipitable water vapor (PWV) is an important indicator to characterize the spatial and temporal variability of vapor. A high resolution atmospheric precipitable can be obtained using ground-based GNSS, but its inversion accuracy usually limited by weighted mean temperature, Tm. For this reason, based on data 17 GNSS stations reanalysis products over 2 years in Hong Kong region, a new model for without Tm parameter established deep learning paper, research results showed that, compared with PWV information calculated traditional parameter, retrieved proposed paper higher, index parameters BIAS, MAE, RMSE are improved 38% average. At same time, was inverted radiosonde study area as reference verify model, it found that BIAS only 0.8 mm, which has accuracy. Further, LSTM more universal when comparable. In addition, order evaluate variation characteristics rainstorm event caused typhoon September 2023, ERA5 GSMaP rainfall were comprehensively used analysis. The show increased sharply arrival occurrence event. After rain stopped, gradually decreased tended stable. have strong correlation extreme events. This shows respond well events, proves feasibility reliability provides method meteorological monitoring weather forecasting.

Language: Английский

---

Prediction method for instrument transformer measurement error: Adaptive decomposition and hybrid deep learning models

Measurement, Journal Year: 2025, Volume and Issue: unknown, P. 117592 - 117592

Published: May 1, 2025

Language: Английский

---

A GRNN-Based Model for ERA5 PWV Adjustment with GNSS Observations Considering Seasonal and Geographic Variations

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(13), P. 2424 - 2424

Published: July 1, 2024

Precipitation water vapor (PWV) is an important parameter in numerical weather forecasting and climate research. However, existing PWV adjustment models lack comprehensive consideration of seasonal geographic factors. This study utilized the General Regression Neural Network (GRNN) algorithm Global Navigation Satellite System (GNSS) China to construct evaluate European Centre for Medium-Range Weather Forecasts (ECMWF) Atmospheric Reanalysis (ERA5) various seasons subregions based on meteorological parameters (GMPW model) non-meteorological (GFPW model). A linear model (GLPW was established accuracy comparison. The results show that: (1) taking GNSS as a reference, Bias root mean square error (RMSE) GLPW, GFPW, GMPW are about 0/1 mm, which better weakens systematic ERA5 PWV. overall Northwest (NWC), North (NC), Tibetan Plateau (TP), South (SC) approximately 0 mm after adjustment. adjusted RMSE four 0.81/0.71/0.62 1.15/0.95/0.77 1.66/1.26/1.05 2.11/1.35/0.96 respectively. (2) three tested using PWV, not involved modeling. 0.89/0.85/0.83 1.61/1.58/1.27 2.11/1.75/1.68 3.65/2.48/1.79 As result, GFPW have adjusting than GLPW. Therefore, can effectively contribute monitoring integration multiple datasets.

Language: Английский

---