Mô hình hóa chất lượng nước biển tại vịnh rạch giá của việt nam, sử dụng ảnh viễn thám sentinel-2 được xử lý trên nền tảng google earth engine

0 Người đánh giá. Xếp hạng trung bình 0

Tác giả: Thanh Sơn Lê, Thị Thu Hằng Lê, Hồng Quảng Nguyễn, Nguyên Minh Nguyễn, Trần Điện Nguyễn, Anh Tuân Vũ

Ngôn ngữ: vie

Ký hiệu phân loại: 526 Mathematical geography

Thông tin xuất bản: Khoa học và công nghệ - đại học thái nguyên, 2022

Mô tả vật lý: 86-96

Bộ sưu tập: Metadata

ID: 489221

Thêm vào giỏ Liên kết toàn văn

Tóm tắt
Chủ đề

Việt Nam có bờ biển dài 3260 km, nơi chất lượng nước biển đang bị ảnh hưởng bởi các hoạt động nông nghiệp, nuôi trồng thủy sản, đô thị hóa, công nghiệp và du lịch, làm dấy lên lo ngại về hệ sinh thái và kinh tế - xã hội ven biển. Vậy nên, việc quan trắc môi trường nước biển ven bờ là rất cần thiết. Nghiên cứu này khai thác khả năng của mô hình học máy vector hỗ trợ để trích xuất nhiệt độ bề mặt nước biển, tổng chất rắn lơ lửng, nhu cầu oxy hóa học và chất diệp lục-a từ ảnh vệ tinh Sentinel-2, được xử lý trong nền tảng đám mây Google Earth Engine. Mô hình máy vector hỗ trợ đã được đào tạo và kiểm định sử dụng dữ liệu trạm đo và từ Cơ quan Khí quyển và Đại dương Quốc gia Hoa Kỳ. Đánh giá độ chính xác mô hình cho thấy có sự tương quan cao giữa dữ liệu thực đo và dữ liệu mô hình hóa khi giá trị R2 trung bình lớn hơn 0,8. Bản đồ phân bố chất lượng nước cho thấy chất lượng nước kém hơn ở khu vực gần thành phố Rạch Giá và các cửa sông, và chất lượng nước tốt hơn ở ngoài khơi. Chúng tôi khuyến nghị sử dụng phương pháp nghiên cứu này và ảnh vệ tinh Sentinel-2 cho các nghiên cứu về chất lượng nước.Vietnam has a 3260 km coastline where the seawater quality is affected by activities of agriculture, aquaculture, urbanization, industry, and tourism, raising concerns on coastal ecosystem and social-economic. Hence, monitoring the seawater environment near the coastline is urgently needed. This research exploits the abilities of the support vector machine model to extract surface seawater temperature, total suspended solids, chemical oxygen demand and chlorophyll-a, from Sentinel-2 images processed in the Google Earth Engine platform. The support vector machine model was trained and validated using in-situ measured data and chlorophyll-a data from the National Oceanic and Atmospheric Administration. Model evaluations showed a strong agreement between the modelled and the training data with a mean R2 greater than 0.8. The water quality distributions presented poorer water quality near the Rach Gia city and river mouths but improved offshore. We strongly recommend this study method and Sentinel-2 data for water quality studies.

Tạo bộ sưu tập với mã QR