Sometimes, studying the variation is the interesting thing

By Andrew L. Rypel

As scientists, we’re trained to key in on ‘response variables’. In my case, fisheries scientists often examine how fish physiology, populations, communities or whole ecosystems react to various environmental drivers or human alteration. Unfortunately, variation in data is too frequently looked upon as a nuisance, an after thought, or worse – a statistical hurdle distracting from presenting the cleanest possible pattern. Yet, what if the variation within the data was the interesting thing all along? Ecosystems are messy and dynamic, but in highly interesting ways. I continually return to this theme, and given that management is often a process of making decisions in the face of high uncertainty, studying variation on its own is probably worthwhile at some level.

Ecologists have long-recognized that understanding spatial and temporal heterogeneity in ecosystems is important. For example, studying spatial distributions of species is extremely common – some are clumped, others are more even, some are random. And population biologists have long-tracked temporal variation in population abundances. Classically, ecologists have linked variability in population numbers to the overall stability of populations and ecosystems (Pimm 1992; Fischer et al. 2001; Zhao et al. 2020). However, understanding relationships between spatial and temporal variability in ecological variables has been more difficult to unravel.

Now at this point, you might ask: Why would understanding such relationships matter? There are of course multiple answers, but at the most basic level, understanding variation is critical to predicting the behavior of dynamic systems. We know from theory and some empirical work in ecosystem and community ecology that spatial heterogeneity may predict temporal heterogeneity (Collins et al. 2018). In one empirical case, spatial variability in plankton production (greenness) in small lakes preceded temporal variation in plankton production (a plankton bloom) (Butitta et al. 2017). Thus, spatial variation may function as an early warning indicator or signal for regime shifts more generally (Nijp et al. 2019).

One of the original challenges in ecology was that replicated spatiotemporal data were exceedingly rare. This limitation is starting to change. Standard, replicated ecological studies and research programs have increased tremendously over the last 30 years. These include massive projects like the Long-Term Ecological Research (LTER) Program sponsored by NSF. Or the newer and more standardized NEON Project, also from NSF. In California, we have access to tremendous long-term monitoring data like the Fall Midwater Trawl Survey from CDFW, the CalCOFI survey, the Delta Juvenile Fish Monitoring Program from USFWS, the Suisun Marsh study started by Peter Moyle at UC Davis, and many others.

In a recent paper (Rypel 2021), I leveraged two unique but distinct datasets to examine spatial and temporal heterogeneity for fish population abundances in natural lakes in northern Wisconsin. Work at this site was started by colleagues at the University of Wisconsin, Center for Limnology as part of the North-Temperate Lakes LTER. Fish populations in these lakes have now been studied consistently for over 40 years – it is an incredible and underutilized data resource. A few key results emerged from the analysis:

Spatial and temporal variation is fish abundances is extraordinary. Populations regularly fluctuate by orders of magnitude within the same lake and across lakes. This is simply interesting – think about the effects of the human population in a given town going from 1,000 to 10,000 in a year or two. As for the spatial variation – this is expected, but it calls back to age-old questions as to why there are more or less animals that live in one ecosystem versus another (Hutchinson 1959).

Fig. 1. Bootstrapped temporal heterogeneity measures for 18 fish species in each of four lakes compared to bootstrapped spatial heterogeneity measures for the same species across 55 lakes regionally. Species like walleye were less heterogeneous in individual lakes over time relative to patterns observed spatially. Conversely, black crappie and yellow perch had high temporal heterogeneity in lakes relative to observed spatial heterogeneity.

The ratio of spatial to temporal heterogeneity is intriguing. Some species, like black crappie and yellow perch showed massive variations within individual lakes, but less variation across lakes. Other species like walleye showed huge spatial variation across lakes, but less variation within lakes over time (Fig. 1). These ratios likely say something about how different species are best managed.  For example, perhaps black crappie are best managed at a local level where common crappie habitats can be manipulated (coarse woody debris, water clarity etc). In contrast, walleye may require more of a landscape-scale strategy involving e.g., enhanced hydrologic connectivity or coordinating mitigation of large-scale impacts to lake riparian environments.

There is a strong relationship between the spatial and temporal heterogeneity. However, it was asymptotic in shape (Fig 2). Empirical studies exploring the relationship between spatial and temporal heterogeneity in ecological variables are scant. Results from this study highlight that spatial heterogeneity in fish abundances predicts temporal heterogeneity, but only at low levels of heterogeneity. Future work could build on these results by testing the generality of this pattern across different ecosystems, taxa, and life-history types.

Fig. 2. Relationship between spatial and temporal heterogeneity in abundance of fishes from north temperate lakes. Each bubble represents a single species, and size of the bubble scales to the number of lakes with temporal data (and thus confidence of temporal heterogeneity patterns). Dark line denotes a weighted non-linear (asymptotic) regression, and weighting was based on the number of lakes used in the mean temporal heterogeneity calculation.

One theme that emerged in this work, and others (Magnuson 1990), is that long-term ecological data are exceptionally important. In general, slow change tends to evade our senses, and if unaccounted for, can lead to blocked understanding. Long-term data are necessary for disentangling short-term variations in ecosystems from long-term trends.

I have been excited by the increase in outstanding synthetic ecological data work coming out of the California water community over the last several years (Stompe et al. 2020; Goertler et al. 2021; Mahardja et al. 2021). There is also new emphasis on boosting access to open data and to data and code provenance. Further exploration of our long-term open access data resources will have the potential to reveal insights on ecology and management. While we do, let’s consider the potential opportunities for describing and studying heterogeneity patterns. Developing an understanding for ecological variation may be critical to uncovering new ideas for protecting California’s native and declining biodiversity.

Further Reading

De La Rosa, G. 2021. Space & Time: Data that push the boundaries of ecology.

Butitta, V. L., S. R. Carpenter, L. C. Loken, M. L. Pace, and E. H. Stanley. 2017. Spatial early warning signals in a lake manipulation. Ecosphere 8(10):e01941.

Collins, S. L., M. L. Avolio, C. Gries, L. M. Hallett, S. E. Koerner, K. J. La Pierre, A. L. Rypel, E. R. Sokol, S. B. Fey, and D. F. Flynn. 2018. Temporal heterogeneity increases with spatial heterogeneity in ecological communities. Ecology 99(4):858-865.

Fischer, J. M., T. M. Frost, and A. R. Ives. 2001. Compensatory dynamics in zooplankton community responses to acidification: measurement and mechanisms. Ecological Applications 11(4):1060-1072.

Goertler, P., B. Mahardja, and T. Sommer. 2021. Striped bass (Morone saxatilis) migration timing driven by estuary outflow and sea surface temperature in the San Francisco Bay-Delta, California. Scientific reports 11(1):1-11.

Hutchinson, G. E. 1959. Homage to Santa Rosalia or why are there so many kinds of animals? The American Naturalist 93(870):145-159.

Magnuson, J. J. 1990. Long-term ecological research and the invisible present. BioScience 40(7):495-501.

Mahardja, B., V. Tobias, S. Khanna, L. Mitchell, P. Lehman, T. Sommer, L. Brown, S. Culberson, and J. L. Conrad. 2021. Resistance and resilience of pelagic and littoral fishes to drought in the San Francisco Estuary. Ecological Applications 31(2):e02243.

Nijp, J. J., A. J. Temme, G. A. van Voorn, L. Kooistra, G. M. Hengeveld, M. B. Soons, A. J. Teuling, and J. Wallinga. 2019. Spatial early warning signals for impending regime shifts: A practical framework for application in real‐world landscapes. Global change biology 25(6):1905-1921.

Pimm, S. L. 1992. The Balance of Nature? Ecological Issues in the Conservation of Species and Communities. University of Chicago Press.

Rose, K. C., R. A. Graves, W. D. Hansen, B. J. Harvey, J. Qiu, S. A. Wood, C. Ziter, and M. G. Turner. 2017. Historical foundations and future directions in macrosystems ecology. Ecology Letters 20(2):147-157.

Rypel, A. L. 2021. Spatial versus temporal heterogeneity in abundance of fishes in north-temperate lakes. Fundamental and Applied Limnology:Published Online,

Stompe, D. K., P. B. Moyle, A. Kruger, and J. R. Durand. 2020. Comparing and integrating fish surveys in the San Francisco Estuary: why diverse long-term monitoring programs are important. San Francisco Estuary and Watershed Science 18(2).

Zhao, L., S. Wang, L. M. Hallett, A. L. Rypel, L. W. Sheppard, M. C. Castorani, L. G. Shoemaker, K. L. Cottingham, K. Suding, and D. C. Reuman. 2020. A new variance ratio metric to detect the timescale of compensatory dynamics. Ecosphere 11(5):e03114.

About Andrew Rypel

Andrew L. Rypel is a Professor and the Peter B. Moyle and California Trout Chair of coldwater fish ecology at the University of California, Davis. He is a faculty member in the Department of Wildlife, Fish & Conservation Biology and Director of the Center for Watershed Sciences.
This entry was posted in Uncategorized and tagged , , , , . Bookmark the permalink.

Leave a Reply