Protected Areas from Space Map Browser with Fast Visualization and Analytical Operations on the Fly. Characterizing Statistical Uncertainties and Balancing Them with Visual Perception

Joan Masό,Xavier Pons,Alaitz Zabala

doi:10.3390/ijgi9050300

Abstract

Despite huge progress in applying Earth Observation (EO) satellite data to protected areas, managers still lack the right tools or skills to analyze the data and extract the necessary knowledge. In this paper a set of EO products are organized in a visualization and analysis map browser that lowers usage barriers and provides functionalities comparable to raster-based GIS. Normally, web map servers provide maps as pictorial representations at screen resolution. The proposal is to use binary arrays with actual values, empowering the JavaScript web client to operate with the data in many ways. Thanks to this approach, the user can analyze big data by performing queries and spatial filters, changing image contrast or color palettes or creating histograms, time series profiles and complex calculations. Since the analysis is made at screen resolution, it minimizes bandwidth while maintaining visual quality. The paper explores the limitations of the approach and quantifies the statistical validity of some resampling methods that provide different visual perceptions. The results demonstrate that the methods known for having good visual perception, the mode for categorical values and the median for continuous values, have admissible statistical uncertainties.

Highlights

Protected Areas (PA) are dynamic systems that need to be studied and managed to preserve the ecosystem services they provide
We propose a solution consisting of implementing a modern web client that is able to combine AJAX, binary arrays, canvas API and the Open Geospatial Consortium (OGC) Web Map Service (WMS) protocol to manage Remote sensing (RS) data structured in coverages
This study considers the impact that simplification has on both visualization and the statistical uncertainties induced when an analysis is made at screen resolution

Summary

Introduction

Protected Areas (PA) are dynamic systems that need to be studied and managed to preserve the ecosystem services they provide. Reflectance cannot be directly interpreted and must be converted into a set of higher-level products estimating comparable variables or are proxies for the values measured with sensors on the ground. Some of these products are directly related to vegetation characteristics, such as leaf area index, gross primary production, vegetation wetness and height, tree cover density, vegetation composition, and classification, among others. The raw data is received at regular intervals and the products can be computed to obtain long time series of homogeneous variables. Examples of products that are only possible with the availability of time series are forest disturbances, phenology, fire occurrence, hydroperiod, etc

Objectives

Methods

Results

Discussion

Conclusion