top of page

First R&D Study of 2024: Validating Satellite Imagery to Detect Poor Water Quality

Updated: Feb 20

By Dr. Paul Monaghan, Fish Welfare Initiative Research & Development Lead


Summary: In order to increase the cost effectiveness of our farmer program (the ARA), we are considering using satellite imagery to detect water quality issues remotely, which would lessen our current reliance on visiting farms individually. This post discusses a study that we began this week, and will run with 20 partner farms for the next five weeks, to validate whether this technology will indeed work for our purposes.


FWI is investing in more and more rigorous R&D this year than previously. We will be running several studies throughout the course of the year in order to identify improved welfare improvements, and will keep our blog updated with our progress. This post is the first of these planned studies—stay tuned for more!


Introduction

We made an important decision in 2023 to establish a new department: the Research & Development (R&D) Department. The stated objective of our R&D Department is to develop new theories of change and to conduct studies to inform FWI programs to improve the lives of fishes. In designing studies to inform new or improved programs, the R&D Department considers the potential for scale from the outset, along with perceived level of impact. 


This post describes one such study that the R&D Department has just begun.


Background

Fish Welfare Initiative’s (FWI) current core program is the Alliance for Responsible Aquaculture (ARA). This program centers on our ground teams collecting water quality data from fish farms and providing the farmers with corrective actions in the event of key water quality parameters indicating that fishes may be exposed to poor conditions. ARA farms also commit to a density cap of 3000 fishes per acre. The ARA currently supports approximately 100 farms primarily in the Kolleru and Nellore regions of Andhra Pradesh, India. The current ARA model requires FWI data collectors to physically visit farms, with the current strategy being to conduct visits once a month to each farm. 


We are currently exploring ways to increase the cost-effectiveness, scalability, and impact of the ARA program:


  • The ARA team is currently exploring potential mechanisms of boosting the number of visits made to farms with the worst water quality issues (see our recent post from January in which we shared a report of a study we conducted at some of the farms with the worst record of water quality). 

  • Simultaneously, the R&D Department is exploring if using satellite imagery to remotely monitor water quality is a viable option, allowing the ARA to collect water quality data more regularly while simultaneously limiting the requirement to actually visit farms.

The latter of these—using satellite imagery to detect water quality issues—is the focus of our new study.


Hypothesis

If our study reveals that key water quality data collected through analysis of satellite imagery are sufficiently accurate and reliable, the ARA model could be modified to exploit remote data collection, allowing for more frequent collection of water quality data at all farms without the need for additional human resources. 



A satellite image view of SNR2, one of the farms participating in this study. We will be comparing assessments generated from imagery like this to the ground measurements our team collects.

Research Question

Is data collected via analysis of satellite imagery sufficiently accurate and reliable to inform the ARA’s decision-making regarding fish farms that require support to improve water quality and corrective actions to recommend to farmers?


Justification for Study

If we can show that using satellite imagery is a viable method to detect farms with high levels of phytoplankton—and/or any other key water quality parameter that can be detected using satellite imagery—we could:


  • Scale up the ARA to support more fish farms with less reliance on “on site” water quality monitoring by data collectors.

  • Collect data from each farm more frequently than currently practiced as part of the ARA.

  • Allow us to offer more targeted support to ARA farmers most at-risk (i.e. those farms that exhibit consistent water quality problems).

  • Improve our approach for identifying future partner farms, for instance by prioritizing regions or farms with known and consistent water quality issues.

If this approach is shown to be valid, it could improve the ARA’s scalability, cost-effectiveness, and impact, helping us to prioritize our resources for farms where water quality issues exist, and thereby improve the lives of fishes most in need.


Study Plans

Yesterday, on February 19, the R&D Department initiated a study to provide evidence as to whether water quality parameters determined through analysis of satellite imagery are sufficiently accurate and reliable to inform decisions on the ground. The results of this study will be used to inform future strategies for possibly scaling the ARA.


Our Data Collector, Durga Prasad, measuring water quality yesterday morning (February 19) at one of the 20 farms in the study.

FWI has no prior experience using satellite imagery, and to test this approach we are partnering with a private partner in India to leverage their experience with satellite imagery. This study will compare water quality data obtained from analysis of satellite images (provided by our partner) with empirical water quality data obtained by direct measurements at the same farms (provided by FWI).


This week we have begun collecting water quality data—via both analysis of satellite imagery and direct measurements—at 20 fish farms supported by the ARA. This will continue over the next five weeks. Afterwards, the data sets will be compared to understand if the water quality parameters determined via analysis of satellite imagery are sufficiently accurate and reliable. 


If the results are positive, we’ll determine how we can incorporate this technology into our programming, and what FWI needs to do; for example, we’ll likely need to build-up in-house capacity to collect and analyze satellite images. But first, in keeping with our commitment to evidence-based decision making, we’re waiting for the results of this study before deciding on how we can leverage it.


We’re excited to commence this study, and we look forward to sharing the results later in the year.

378 views0 comments
bottom of page