Social sensor as an emerging remote sensing tool and its application to sensing taal lake fishkill

Issue Date

2021

Abstract

Fishkill in Taal Lake has long been a socio-economic problem of the people who rely on the lake for livelihood. More recently, the national government considered fishkills as a food security concern. Although recorded occurrences of fishkill show seasonality patterns, predicting with high fidelity when within seven days and where in the vast expanse of the lake will fishkill occur is still a challenge. A fishkill early warning system (FEWS) based on predictive models requires an expensive three-dimensional array of physico-chemical sensors that must not only span the lake surface but must also extend down within the fish culture depth. Currently, such a network of sensors is an impractical approach with a projected annual operational and maintenance cost greater than the amount it will save to avoid a fishkill losses in a year. In recent years, the social media has become one of the main avenues for expressing real-time concerns among technology-aware Filipinos. Because the social media has attained unprecedented reach even in the remote areas of the Philippines, it is seen as a possible cheaper alternative to sensor-dependent FEWS through what is called "social sensors." Social sensors are the "tweeted" concerns of lake residents when fishkill symptoms have been observed within the lake. In this study, an ICT-based system was constructed to collect dated and geolocated microposts from Twitter users within the 10-km radius of Taal Volcano (14°N, 121°E). The system collected tweets from February 2013 to January 2014 and computed the tweets' respective overall contextual sentiment polarities (OCSP). Two fishkill events were reported within this time period. Data show that neutral to positive OCSP prevail when no fishkills have been reported, while increasing frequency of negative OCSP's was observed 1-7 days before the reported fishkills. The co-occurrence of prevailing negative OCSP and fishkill event may provide the much sought-for fidelity to "nowcast" fishkill at a much cheaper price, or to augment existing physico-chemical-based predictive models. Such an ICT-based system may be customized to cheaply "social sense" other natural phenomena that threatens food security in the Asian region. The utility of this system to improve the nowcasting capability of FEWS underscore the emergence of a new remote sensing tool that can augment, if not replace, current conventional physico-chemical-based sensors.

Source or Periodical Title

ACRS 2015 - 36th Asian Conference on Remote Sensing: Fostering Resilient Growth in Asia, Proceedings

Page

1-8

Document Type

Conference Paper

Physical Description

graphs

Language

English

Subject

Emerging remote sensing tool, Social sensor, Taal lake fishkill, Twitter

Digital Copy

yes

Link to Full Text

Share

COinS