Hypoxic Floods and Fish Kills

Our latest paper was just published in Nature Communications!

Organic matter loading by hippopotami causes subsidy overload resulting in downstream hypoxia and fish kills

From the first fish kill we witnessed almost a decade ago in the Mara River, we have painstakingly been tracing the fish kills and hypoxic floods to the flushing of hippo pools.  We're happy to get this paper in an open access journal so that anyone can access it anywhere.  

Here are some popular press pieces about this work:

• https://www.theatlantic.com/science/archive/2018/05/hippos-poop-so-much-that-sometimes-all-the-fish-die/560486/
• http://www.sciencemag.org/news/2018/05/excess-hippo-dung-may-be-harming-aquatic-species-across-africa
• http://www.cbc.ca/radio/quirks/may-19-2018-psychadelic-science-understanding-kilauea-up-close-hippo-poop-kills-fish-1.4667264/flushing-hippo-toilets-smother-fish-in-african-rivers-1.4667344

First drowning of the season

We recently published 5 years of research on wildebeest mass drownings in the Mara River. We plan to continue studying this phenomenon for the foreseeable future, both to document the occurrence and size of the drownings and to continue understanding their influence on the river ecosystem. On July 19, 500 wildebeest were crossing the Mara River near the lower bridge, and 150 drowned. This is a relatively small drowning compared to some others we have seen and compared to the annual mean of 6,250 we have documented over the years. However, we're still early in the migration season, and the river has been unseasonably low, making it easy for the wildebeest to cross without incident. So, we'll see... We're in the field until the end of the migration in November, so we'll keep count.

Car parts and cappuccinos

Often an errand that should take an hour or two ends up taking half a day. Rarely is that half day spent so enjoyably as having a cappuccino overlooking the incomparable view at Angama, a new camp perched over the edge of the Mara Triangle. When we arrived at their airstrip to pick up our car parts and found out we'd been given the wrong time and would have to wait several hours, Tyler, the manager at Angama, graciously let us wait at the lodge and enjoy a cappuccino. While there, we met a guest from NYC who had just read about our research on wildebeest mass drownings. What started as a short errand turned into a lovely morning! One of the joys, and frustrations, of working here is never knowing how each day will end, but this time it was definitely a joy!

450 lbs. of gear

We headed back to Kenya in June for a 6 month field season. We had to purchase a plane ticket for Lily this year, since she's now over 2. The one upside... 3 extra bags on her ticket! Of course, about 2 of those bags were filled with her field wardrobe and clothes, but we managed to get an extra field bag in there!

The Mze

After 8 years of faithful service (okay, maybe not always faithful, but certainly adventurous!), our old, white Land Rover was finally due for a new coat of paint. Here she is, looking pretty sharp in her new dark green coat. Now it's difficult to tell this one apart from our newer Land Rover purchased last year, which is also green, so we've taken to calling this one the Mze (the elder) and the newer one the Kijana (the youth).

Publishing our Research on Wildebeest Mass Drownings

This week our scientific paper on wildebeest mass drownings in the Mara River was published in the Proceedings of the National Academy of Sciences. It is so exciting to finally publish this research, which essentially began with our first observations of wildebeest carcasses in the Mara River in 2008. Over the years, we have spent hundreds of hours (at least!) observing wildebeest crossings, counting carcasses, collecting water and tissue samples, measuring nutrient chemistry and stable isotope signatures, analyzing data, and writing up this manuscript. After completing the paper, it then went through a rigorous peer review process and months of edits and additional data analyses. Throughout this process, our understanding of the scale of these drowning events and their ecosystem effects has continued to evolve.

We have now shown that mass drownings occurred in 13 of the last 15 years in the Mara, bringing an average of 6,250 carcasses into the river each year. These carcasses equal 1100 tons of biomass, which contributes 107 tons of carbon, 25 tons of nitrogen, and 13 tons of phosphorus to the river each year. About half of the carcass is soft tissue, which decomposes within weeks to months. The other half is bone, which decomposes over 7.5 years, slowly leaching nutrients, especially phosphorus, into the river. The carcasses alter nutrient cycling in the river and provide important resources for a range of terrestrial and aquatic scavengers. When carcasses are present, they comprise up to 50% of the diet of fish in the river. Even months after carcasses have decomposed, the biofilm that grows on bones continues to be an important food resource.

This paper represents our efforts to understand the natural phenomenon of mass drownings in one of the few places left on the earth where you can study the impacts of large animal migrations on unregulated rivers, and it helps give us a window into what other rivers may have looked like when large animal migrations and unregulated rivers were still common features of our landscapes. It also charts a course forward for us in terms of the next set of questions to investigate. As we take a moment to reflect on all that has gone into the development of this paper, I thought it would be interesting to look back at some of our old blog posts as we studied these events.

The first time I witnessed a drowning happening

Some videos of avian scavengers using carcasses 

Chris's observations of a huge drowning happening

Early thoughts on bones in the river

The Dry Season

We're heading into the dry season in the Mara, which generally runs from January until early-mid March when the long rains come. Geemi just sent us some pictures of the Mara River, showing how low the river is. You can actually see the river level on the real-time river gauge on the top right of our blog (under Mara River Water Level). Chris built this river gauge using low-cost Arduino sensors, and it uploads real-time data on the river every 15 minutes. When the arrow is in the red zone, the river is lower than the minimum sustainable flow levels that were identified in the Environmental Flow Assessment for the Mara. You can tell from looking at these pictures that the river is well below where it normally is, and the dry season is just getting started. Hopefully we'll get a few big rains to help fill things up before we head into February...

Looking downstream into Tanzania from the Purungat Bridge

Looking upstream from the Purungat Bridge

Kwaheri Mara!

It's always bittersweet leaving the Mara. This place is our second home, and we have a beautiful camp, good friends and an amazing river we hate to leave behind. On the other hand, after several months of no running water, limited solar power, and the same 4 outfits, you do start to long for modern conveniences. I usually crave salad, hot showers, a nice cappuccino, and a leisurely morning spent on my computer with the screen at full brightness. Nevertheless, the time has come to return to the US-- we have a lot of samples to analyze in the lab, lots of papers to work on writing, and Chris is a teaching assistant for a class this semester. It's going to be an exciting year... We'll be working with Ella and James to write up their senior research projects. Chris and I both have several papers we're working on publishing from our research. And we have a super exciting field season we're planning for 2017! Already excited to think about heading back here.... Until then, kwaheri Mara (goodbye)!

Hippos on the bank of the Mara River

Our home in the Mara

Why we need wiper blades on our sensors

It's really important for our research to be able to measure water quality in the river on a continual basis over long periods of time. To do this, we have some pretty amazing water quality sensors made by Eureka, that can measure lots of different water quality variables and store the data for weeks or months at a time. We have had these sondes for 6 years, and we have put them into some of the most challenging conditions I can imagine, and they just keep plugging along. Our model has the extra benefit of a wiper blade that cleans the surface of each sensor just before taking a new measurement, which is critical in the Mara for preventing biofilm or dirt from fouling the sensor and compromising the measurement. However, we still have to pull the sonde out of the river every few weeks just to clean off all the bugs, hippo poop and biofilm that accumulates.

Our water quality sonde after 3 weeks in the river

Where is all the algae going?

Hippos and wildebeest load a lot of nutrients into the Mara River... like, a lot. All of those nutrients should fertilize the river and produce a lot of algae. It's what the textbooks say should happen, and it's what research in other rivers show should happen. We see it happen in some portions of the river where there are only moderate levels of hippo inputs. However, it is not what happens where the hippo and wildebeest inputs are highest. Why not? I don't know. 

This-- getting results that don't fit your expectations, that surprise you-- is simultaneously the most exciting and the most frustrating part of science. It's the part that makes science addictive, that keeps you coming back for more, that makes you think grand discovery is on your doorstep. It's also the part that can drive you crazy and make it difficult to publish your research. If you're getting an unexpected result, you need to 1) be absolutely sure it's a real result and not an artifact of something you screwed up on accident, and 2) figure out why you're getting it. "This is weird, and I don't know why it's happening," will only get you so far. 

So, back to our algae... Why don't we see more in the river? Is it not growing in the first place? Or is it growing and going somewhere? I have been asking this question a lot over the last year as I have been working on the final papers from my dissertation, and this summer I was able to run a pilot experiment to test what is becoming my primary hypothesis... hippo and wildebeest inputs provide so much food for bugs and fish that their populations increase and they then graze down the algae. This would be really cool if it turns out to be true, because it would show that animal inputs can enter the river food web through various pathways, and which pathway they follow determines how those inputs shape the river ecosystem. Or I could be wrong and it could be something else entirely... See? Exciting, and a little frustrating. 

So, to test this hypothesis, I set out a grazing exclusion experiment this summer. I basically grew algae in the river on glass discs (the white circles in the picture below). Some discs were protected from grazing (either by a cage to restrict fish or by insecticide to restrict bugs) and others were not. If there was more algae growing on the protected discs than on the not-protected discs, that would suggest that algae can grow in the river, but it doesn't accumulate there because it gets grazed down. 

Before deployment in the river

I wasn't originally planning to run this experiment this year, as we already had a full schedule, but I was excited to be able to put this together almost completely with supplies we had stored around our camp. I built two of these that I deployed at two different sites in the river. I left them out for two weeks and hoped they didn't get stepped on by a hippo, washed away in a flood or collected by an enterprising Maasai person with a need for some glass discs. I always consider it a major win when we put something out in the river and return later to find it still there. 

After 2 weeks in the river

After two weeks, we collected the glass discs and measured the chlorophyll on them as an indication of how much algae was growing there. To my excitement, the discs protected from grazing by insects had more algae than the discs that weren't protected, supporting my hypothesis! To my frustration, the cages protecting the discs from fish caught a lot of hippo feces which influenced those data, and the variability was so high that my insect results weren't statistically significant. Ah science... the perfect career for those who love a good challenge! So, I can't say I've solved this mystery yet, but I'm getting closer, and I'm already excited about getting back to the field next year to try again.