SUPERVISED CLASSIFICATION METHODS FOR AUTOMATIC DAMAGE DETECTION CAUSED BY HEAVY RAINFALL USING MULTITEMPORAL HIGH RESOLUTION OPTICAL IMAGERY AND AUXILIARY DATA

Abstract. In the context of climate change and rising frequency of extreme hydro-meteorological events around the world, flood risk management and mapping of heavy rainfall-related damages represent an ongoing critical challenge. For decades now, remote sensing has been largely used to investigate spatial and temporal changes in land use and water resources. Today, different satellite products provide fast and crucial knowledge for the study of hydrological disasters over large areas, possibly in remote regions, with high spatial resolution and high revisit frequency. Yet, until now, few works have sought to detect the full range of extreme rainfall-related damages with optical imagery, especially those caused by intense rainwater runoff beyond the direct vicinity of major waterways. The work presented in this paper focuses on the Aude severe weather event of October 15th, 2018, in the South of France, for which more than a thousand claims for agricultural disaster were registered, both related to river overflowing and rainwater runoff.The full resources of ground truths, contextual information, land use as well as digital elevation model (DEM) combined to high resolution and high frequency optical imagery (Sentinel-2, Pléiades) are used to develop an automatic damage detection method based on supervised classification algorithms. Through the combination of several indicators characterizing heterogeneous spectral variations among agricultural plots following the event, a Gaussian process classifier achieved various classification accuracies up to 90% on a large comparable and independent photo-interpreted validation sample. This work builds great expectations for applications in other areas with contrasted climate, topography and land cover.

Download Full-text

Urban land use mapping using high resolution SAR data based on density analysis and contextual information

Canadian Journal of Remote Sensing ◽

10.5589/m13-002 ◽

2013 ◽

Vol 38 (6) ◽

pp. 738-749 ◽

Cited By ~ 13

Author(s):

Zhaohua Chen ◽

Ying Zhang ◽

Bert Guindon ◽

Thomas Esch ◽

Achim Roth ◽

...

Keyword(s):

Land Use ◽

High Resolution ◽

Contextual Information ◽

Urban Land ◽

Urban Land Use ◽

Density Analysis ◽

Sar Data

Download Full-text

Detailed Lacustrine Calving Iceberg Inventory from Very High Resolution Optical Imagery and Object-Based Image Analysis

Remote Sensing ◽

10.3390/rs12111807 ◽

2020 ◽

Vol 12 (11) ◽

pp. 1807

Author(s):

Julian Podgórski ◽

Michał Pętlicki

Keyword(s):

High Resolution ◽

Contextual Information ◽

Tidewater Glaciers ◽

Object Based Image Analysis ◽

Object Based ◽

Volume Relationship ◽

Optical Imagery ◽

Very High ◽

Classification Product ◽

Area And Volume

In the field of iceberg and glacier calving studies, it is important to collect comprehensive datasets of populations of icebergs. Particularly, calving of lake-terminating glaciers has been understudied. The aim of this work is to present an object-based method of iceberg detection and to create an inventory of icebergs located in a proglacial lagoon of San Quintín glacier, Northern Patagonia Icefield, Chile. This dataset is created using high-resolution WorldView-2 imagery and a derived DEM. We use it to briefly discuss the iceberg size distribution and area–volume scaling. Segmentation of the multispectral imagery produced a map of objects, which were classified with use of Random Forest supervised classification algorithm. An intermediate classification product was corrected with a ruleset exploiting contextual information and a watershed algorithm that was used to divide multiple touching icebergs into separate objects. Common theoretical heavy-tail statistical distributions were tested as descriptors of iceberg area and volume distributions. Power law models were proposed for the area–volume relationship. The proposed method performed well over the open lake detecting correctly icebergs in all size bands except 5–15 m2 where many icebergs were missed. A section of the lagoon with ice melange was not reliably mapped due to uniformity of the area in the imagery and DEM. The precision of the DEM limited the scaling effort to icebergs taller than 1.7 m and larger than 99 m2, despite the inventory containing icebergs as small as 4 m2. The work demonstrates viability of object-based analysis for lacustrine iceberg detection and shows that the statistical properties of iceberg population at San Quintín glacier match those of populations produced by tidewater glaciers.

Download Full-text

IDENTIFICATION OF EXTREME RAINFALL EVENTS USING HIGH-RESOLUTION CARBON ISOTOPE MEASUREMENTS ACROSS TREE-RINGS

10.1130/abs/2017am-306100 ◽

2017 ◽

Author(s):

Rose G. Telus ◽

◽

Brian A. Schubert

Keyword(s):

High Resolution ◽

Carbon Isotope ◽

Tree Rings ◽

Extreme Rainfall ◽

Extreme Rainfall Events ◽

Rainfall Events ◽

Isotope Measurements

Download Full-text

The Development of a Quantitative Precipitation Forecast Correction Technique Based on Machine Learning for Hydrological Applications

Atmosphere ◽

10.3390/atmos11010111 ◽

2020 ◽

Vol 11 (1) ◽

pp. 111 ◽

Cited By ~ 2

Author(s):

Chul-Min Ko ◽

Yeong Yun Jeong ◽

Young-Mi Lee ◽

Byung-Sik Kim

Keyword(s):

Machine Learning ◽

Heavy Rainfall ◽

Extreme Rainfall ◽

Machine Learning Techniques ◽

Precipitation Forecast ◽

Machine Learning Technique ◽

Rainfall Forecast ◽

Quantitative Precipitation Forecast ◽

Correction Technique ◽

Learning Technique

This study aimed to enhance the accuracy of extreme rainfall forecast, using a machine learning technique for forecasting hydrological impact. In this study, machine learning with XGBoost technique was applied for correcting the quantitative precipitation forecast (QPF) provided by the Korea Meteorological Administration (KMA) to develop a hydrological quantitative precipitation forecast (HQPF) for flood inundation modeling. The performance of machine learning techniques for HQPF production was evaluated with a focus on two cases: one for heavy rainfall events in Seoul and the other for heavy rainfall accompanied by Typhoon Kong-rey (1825). This study calculated the well-known statistical metrics to compare the error derived from QPF-based rainfall and HQPF-based rainfall against the observational data from the four sites. For the heavy rainfall case in Seoul, the mean absolute errors (MAE) of the four sites, i.e., Nowon, Jungnang, Dobong, and Gangnam, were 18.6 mm/3 h, 19.4 mm/3 h, 48.7 mm/3 h, and 19.1 mm/3 h for QPF and 13.6 mm/3 h, 14.2 mm/3 h, 33.3 mm/3 h, and 12.0 mm/3 h for HQPF, respectively. These results clearly indicate that the machine learning technique is able to improve the forecasting performance for localized rainfall. In addition, the HQPF-based rainfall shows better performance in capturing the peak rainfall amount and spatial pattern. Therefore, it is considered that the HQPF can be helpful to improve the accuracy of intense rainfall forecast, which is subsequently beneficial for forecasting floods and their hydrological impacts.

Download Full-text

High-Resolution Climate Change Impact Analysis on Expected Future Water Availability in the Upper Jordan Catchment and the Middle East

Journal of Hydrometeorology ◽

10.1175/jhm-d-13-0153.1 ◽

2014 ◽

Vol 15 (4) ◽

pp. 1517-1531 ◽

Cited By ~ 15

Author(s):

Gerhard Smiatek ◽

Harald Kunstmann ◽

Andreas Heckl

Keyword(s):

Climate Change ◽

High Resolution ◽

Water Availability ◽

Impact Analysis ◽

River Flow ◽

Climate Model ◽

Mesoscale Model ◽

Global Climate Model ◽

Full Range ◽

The Impact

Abstract The impact of climate change on the future water availability of the upper Jordan River (UJR) and its tributaries Dan, Snir, and Hermon located in the eastern Mediterranean is evaluated by a highly resolved distributed approach with the fifth-generation Pennsylvania State University–NCAR Mesoscale Model (MM5) run at 18.6- and 6.2-km resolution offline coupled with the Water Flow and Balance Simulation Model (WaSiM). The MM5 was driven with NCEP reanalysis for 1971–2000 and with Hadley Centre Coupled Model, version 3 (HadCM3), GCM forcings for 1971–2099. Because only one regional–global climate model combination was applied, the results may not give the full range of possible future projections. To describe the Dan spring behavior, the hydrological model was extended by a bypass approach to allow the fast discharge components of the Snir to enter the Dan catchment. Simulation results for the period 1976–2000 reveal that the coupled system was able to reproduce the observed discharge rates in the partially karstic complex terrain to a reasonable extent with the high-resolution 6.2-km meteorological input only. The performed future climate simulations show steadily rising temperatures with 2.2 K above the 1976–2000 mean for the period 2031–60 and 3.5 K for the period 2070–99. Precipitation trends are insignificant until the middle of the century, although a decrease of approximately 12% is simulated. For the end of the century, a reduction in rainfall ranging between 10% and 35% can be expected. Discharge in the UJR is simulated to decrease by 12% until 2060 and by 26% until 2099, both related to the 1976–2000 mean. The discharge decrease is associated with a lower number of high river flow years.

Download Full-text