Dissolved Oxygen Tolerance Study: Findings for Rohu

Summary

In July 2024, we announced the initiation of a research project to determine the lowest level of dissolved oxygen that Indian major carps (IMC) can tolerate without exhibiting welfare concerns. This research was conducted in our test facility in Andhra Pradesh and involved studying fishes exposed to differing levels of oxygen in tanks. We faced a number of challenges during this project, and unfortunately, we lost a number of fishes. Losing fishes caused us to pause and reflect, and consider whether we should proceed (see our August 2024 newsletter). Ultimately, given the importance of this study to our evidence-base and for informing our core farmer program, the Alliance for Responsible Aquaculture (ARA), we made the decision to proceed, albeit with improvements to our processes.

After making improvements to our processes and protocols, in March 2025, we completed a study on rohu, the most prevalent species of IMC farmed in India. The results of this study indicate that 3 mg/L dissolved oxygen appears to be the welfare threshold for rohu, which is in line with the thresholds currently used by our farm engagement program. When we repeated this study for catla, the second most prevalent species of IMC farmed in India, unfortunately, we lost more fishes, which prompted us to terminate the study early. 

Given the number of fishes we lost during this research project, and the stress we would need to expose fishes to, we need to consider if it is acceptable or not for FWI to reinitiate the DO Tolerance study for catla. We have taken time once again to improve protocols, systems, and infrastructure, but we have yet to make a decision as to whether to proceed.

Background

Dissolved oxygen (DO) is considered the key water quality parameter by FWI, and many decisions by FWI around fish welfare are based on DO levels. Indian major carps (IMCs)—the primary fishes targeted by FWI in India—are considered quite tolerant to DO compared to many other fishes, but few studies have been conducted on IMCs to determine their oxygen requirements. Given the sparsity of published data about IMC and how they are or are not impacted by specific water quality parameters—in particular, DO—FWI initiated a research study to address this knowledge gap. More specifically, we conducted experiments to ascertain the lowest level of DO which IMC can tolerate without exhibiting welfare concerns.

Overview of Study

To conduct this study, we established a purpose-built test facility at our former research site at Adikavi Nannaya University (AKNU) in Rajamahendravaram, Andhra Pradesh. This test facility comprised eight tanks in total: six 500-L experimental tanks, which are used for conducting experiments; and two 1500-L stock tanks, which are used for housing fishes before and after experiments.

The DO Tolerance study involves an acclimatization period, where fishes—procured directly from a supplier and transported to the test facility by truck—are transferred directly from the transport truck into the stock tanks. If fishes feed and behave normally, they are transferred to the experimental tanks (24 fishes per tank) and allowed to acclimatize for at least 5 days before initiating the experiment. We used 144 fishes (125 g average size) divided into three groups:

1. Control Group: Oxygen kept stable at 5 mg/L, ammonia kept low. 

   
   

2. Experimental Group: Oxygen gradually reduced from 5→4→3→2→1 mg/L over 48 hours by injecting nitrogen gas. Ammonia increased naturally as oxygen dropped.

   
   

3. Ammonia Control Group: Oxygen kept stable at 5 mg/L, but ammonia was artificially increased to match levels seen in the experimental group. This allowed us to separate any observed welfare effects resulting from increased ammonia (which increases naturally in the experimental group as oxygen levels drop) from effects resulting from decreased oxygen.

   
   

We measured the following parameters:

• Blood Parameters (taken every 12 hours):

  • Hemoglobin, hematocrit, and red blood cells (indicators of oxygen-carrying capacity)

  • Analysis of stress hormones was planned, but it proved difficult to find a reliable lab capable of conducting the necessary tests

    
    

• Behavior (video and real-time observations):

  • Activity level (no movement, slow swimming, or active/erratic)

  • Position in water (bottom, mid-water, or surface)

  • Shoaling behavior (how tightly fishes group together)

  • Gasping at the surface

    
    

• Feeding Behavior (tested 24 hours after stress exposure):

  • How quickly fishes respond to food

  • What percentage of fishes in the tanks ate

  • How vigorously they fed

    
    

At the end of the study, fishes were released into their natural habitat in the wild, as per FWI’s ethical standards. The study was conducted separately for rohu and catla species, the two species of IMC most abundantly farmed in India.

Our test facility where we conducted the DO Tolerance study. The two larger tanks in the foreground (left and right) are the 1500-L stock tanks, with the six 500-L experimental tanks in the background, along with oxygen tanks (to improve oxygen during the experiment).

Study on Rohu

Rohu was selected as the first species to study. Rohu is the predominant carp species farmed in India, constituting approximately 80% of the fishes in a typical IMC farm. Although we announced the study in July 2024, we suffered from an aborted attempt at the study in July, which prompted us to make changes to our facility, systems, and protocols to mitigate the risk of stressing or losing fishes. This took time, but suffice it to say, the welfare of fishes is our priority. Ultimately, we conducted the rohu study between March 5–23, 2025.  

Summary of Key Findings

The study showed that 3 mg/L DO is the welfare threshold for rohu (i.e., the minimum level needed to maintain normal behavioral function and stress coping capacity). While rohu can survive at 2 mg/L DO through compensatory mechanisms, such as increased hematocrit and red blood cell production, this apparent stability masks significant welfare concerns. In short, below 3 mg/L DO, fishes experience:

• Behavioural impairments, including reduced activity, loss of positional flexibility (trapped at the bottom of the tank), and suppressed feeding behaviour. 

  
  

• Lasting effects, with fishes showing impaired feeding even after 24 hours of recovery in optimal conditions. Fishes exposed to low dissolved oxygen showed only 42% feeding engagement vs. 100% in fishes under control conditions, with 8 times slower response times.

  
  

When DO drops, ammonia naturally increases (reaching 0.77 mg/L at 3 mg/L DO). The low DO/high ammonia interaction amplifies welfare concerns, creating a behavioral trap:

• Fishes under adequate DO (5 mg/L) respond to ammonia irritation with hyperactivity, which helps to increase gill ventilation.

  
  

• Fishes under low DO (<3 mg/L) show lethargy at the tank bottom, where ammonia accumulates most, showing that the protective hyperactivity response is completely suppressed, worsening the effects of exposure.

Summary of key findings observed for rohu during the DO tolerance study.

Currently, our Alliance for Responsible Aquaculture (ARA) program issues corrective actions to farmers when DO is outside of FWI’s recommended range of 3-5 mg/L (based on measurements taken in the morning). The lower threshold of 3 mg/L applied by the ARA is supported by the findings from our DO Tolerance study.

You can access our internal report on this study here, an analysis of the data here, and the raw water quality data.

Internal Report

Analysis of Data

Raw Water Quality Data

FWI staff taking water quality measurements during the study.

Loss of Fishes During the Study

Unfortunately, during the course of the rohu DO tolerance study, we lost a number of fishes. In our earliest study, in July 2024, we lost 25 fishes during the acclimatization period before we actually could begin the study. Twenty of these fishes were lost by jumping out of the tank during a period when no team member was present. This prompted us to shut down the experiment and revisit our processes (we released all remaining fishes into their natural environment). One of the important changes we subsequently made was to install netting over the tanks. In the main study (reported above), nine fishes were lost (seven during the acclimatization period before the actual study, and two during the wild release process at the very end), with these losses attributed to transportation stress.

While losing fishes is always upsetting, any study involving animals always comes with the risk of losing some of them. We mitigate that risk as much as we can. For the rohu study (after the initial one we terminated early), we believe that these mitigation measures were appropriate, and that the fishes we lost were not directly due to our actions (although, we are very cognizant that the very nature of involving these fishes in the study led to their deaths).

Study on Catla

Overall, we believe the study with rohu went well—notwithstanding the loss of 25 fishes lost during the initial aborted study in July 2024 and 9 fishes during the principal study in March 2025—and gave us the confidence to proceed with the catla study. The same study, this time focusing on catla, was planned for April 9–25, 2025.

We introduced catlas into the stock tanks on Wednesday, April 9th, to initiate the acclimatization process. Unfortunately, within four days we had lost 24 fishes, likely due to the stress involved in the harvesting/transportation process (i.e. sourcing fishes from the supplier and bringing them to our facility) combined with stress due to higher than normal water temperatures (water temperature at our facility had risen to a higher than ideal level, which we had no control over). Given the losses and our inability to adequately control water temperature, we terminated the study and released the remaining fishes; one additional fish died during the wild release process, and two others, whom we did not expect to survive in the wild, were euthanized, bringing the total losses to 27 fishes in five days.

We did not proceed with conducting the study, but rather took time to reflect on whether the risks associated with these types of experiments are acceptable or not to FWI, and to consider strategies to mitigate against these issues happening in the future, should we decide to proceed with the study again. 

Path Forward

The rohu trial provides valuable evidence that current oxygen management practices in IMC ponds may need to maintain a minimum threshold of 3 mg/L for higher welfare, not just to prevent mortality, but to ensure behavioral welfare and production performance.

The findings from the rohu study support the lower thresholds currently used by our ARA program. Given that rohu and catla are generally farmed together in the same ponds, there is value in determining if those lower DO thresholds also apply to catla. If the lower limit for catla is notably higher (e.g., 5 mg/L), that would suggest that the DO thresholds applied by the ARA should be modified so that the welfare of both species is improved. 

However, given the number of fishes we lost during the study, and the stress we would have exposed all fishes to, we need to consider if it is acceptable or not for FWI to reinitiate the DO Tolerance study for catla. We have taken time to improve protocols, systems, and infrastructure, but with our 2026 Goal in mind, our decision to proceed with this study is dependent on other priorities. We expect to decide by the end of the second quarter of 2026 as to whether we’ll proceed with this study on catla.