The timing of precipitation

Traces of cumulative precipitation (standardized across years) for Konza.  Date range is 145-250 (May 25 - Sep 7), which is the precipitation timing critical climate period.

When rain falls might not seem to be important, but in many ecosystems it is often more important than how much rain falls.

In annual grasslands, the germination of different species if dependent on when rains begin. And the end of the rains set the end of the growing season. Same could be said for many tropical grasslands.

Many deserts are structured by when rain falls as shrubs with deep roots benefit from winter rains and grasses summer rains.

But what about temperate grasslands where the timing growth is dependent more on temperature than rainfall? How important is timing?

Put another way, if climate change causes rains to fall later (or earlier), will it matter.

There has been little work to directly test the importance of timing of rainfall.

I recently dug through the Konza climate data to see how much the timing of precipitation varied among years. Quite a lot. For May-September rainfall, the timing of when the average unit of precipitation fell varied by over six weeks over a 25-y period. In some years, half the rain had already fallen by early June. In other years, it took to mid August.

Does it matter? If plants aren't running out of water and it all goes into the same bucket, how much impact can there be of the bucket filling early or late?

Turns out a lot.

Variation in the timing of rainfall explained as much variation in grass productivity as temperatures and about half as much as the total amount of rain.

From these calculations, shifting precipitation by just 1 week later in the season can reduce productivity by 5-10%. 

For reference, a decrease in productivity of this amount is enough to put most ranchers out of business, all other things equal. 

How to say "It's going to be a hot one"

If there is anywhere where it would be hard to predict the weather, it would have to be here in Kansas. We are just too far from the sea for anything to have strong influence on our weather.

Yet, it seems like it must be a great temptation to try to predict the weather. Because many people have said to me in casual conversation some variation of, "It's going to be a hot summer." Whatever the weather is today (or the past few days) people seem to project forward a couple of months.

Is there any basis for this?

If it's hot this month, is it likely to be hot next month? Or dry?

Here's how you test that.

First you generate climate anomalies. In a seasonal climate like ours, it's simple to know what the climate will be like in a month or 6 months or a year.

If it's June, in the grand scheme of things, it's likely to be hot next month cold in 6 months and hot again in 12 months.

But, what we're interested here is in the climate anomaly and how they are correlated over time.

So we aren't saying it's going to be hot next month, but hotter than average next month.

To evaluate this, you look at past records of data and run autocorrelation analyses. This analysis looks at how different temperature or precipitation over a certain length of time and then examines the correlation between the current period's anomaly (hotter/colder, wetter/drier) vs. next month's.

Here are some data from May weather over the past 50 years for Manhattan KS.

I went to knmi.nl and accessed their climate explorer. The site runs autocorrelations for any weather station in the world.

Here's what the autocorrelation analysis for daily maximum temperatures look like for 30-d periods throughout May and June (start dates of May 1 to May 30).



What is important in this graph is the correlation after 30 d. Essentially it says that the correlation coefficient between the 30-d starting in May and the following 30-d is less than 0.2.

How helpful is that?

Here's data for Konza (close to Manhattan) for 1984-2010, which I've been using for analyses lately.

X-axis is mean maximum temperatures for day of year 120-149 (roughly May) vs. same temperatures in the next 30 d.



Correlation coefficient of this is about 0.3 (similar to autocorrelation analysis).

You can see that there is about 9 °C variation among years in May temperature. If it's the hottest May on record, June is likely to be just 1°C hotter that average. That's not much predictive capacity.

And the 30-d after that? r = 0.2 and no significant predictive capacity. It could be equally likely to be the hottest month on record as the coldest.

For precipitation, it gets even worse.

here's the same autocorrelation analysis for precipitation during that period.

Essentially for the 30-d following May, there is no significant predictive capacity (maybe slightly drier).

 Precipitation between the two 30-d periods at Konza: flat.

As a caveat, i should say that these patterns have nothing to do with general trends over long time periods associated with global warming. Only the structure of seasonal variability.

That said, it's probably more certain to say that the future is going to be hotter than it is to say that the next month is hotter (or colder or wetter or drier).

At least here in Kansas.

Where to send your paper after it's been rejected

Ever wonder where to send your paper after it's been rejected from a journal? 

The authors of a recent Science paper mapped the submission history of over 80,000 scientific articles published from 2006-2008 .

This monumental task allows tracking of the most likely flows of papers through journals. 

The network map provides more than a guide to where to send your next rejected paper.

A few interesting results came out of this. 

First, papers that had been rejected from another journal were cited more. Is this because papers were improved with further revisions? Or is it because papers that are likely to be the most interesting don't often fit into the neat model of a given journal's papers? 

Second, resubmissions often went to journals with lower impact factors. No surprise there, but good to quantify. 

Third, the proportion of papers published that had never been submitted elsewhere was similar across a range of impact factors of journals. The range was really narrow in the grand scheme of things.  

Publishing is tricky. There is an immense amount of wasted effort trying to get papers published. Given that, journals and authors are reasonably efficient in many respects. 

Flows of Research Manuscripts Among Scientific Journals Reveal Hidden Submission Patterns

V. Calcagno et al. Science 338, 1065 (2012)

DOI: 10.1126/science.1227833

Planetary boundary for nitrogen

A paper a few years ago (Rockström et al Nature 2009) laid out a conceptual framework for how humans are impacting the earth ecosystem. They identified nine ways that humans impact the Earth and  identified thresholds at the global scale (planetary boundaries) that once exceeded could have severe consequences for humans.

They stated that three boundaries have already been exceeded. The rate of climate change and biodiversity loss were the first two. The third was biogeochemical flows, specifically the N cycle.

Direct and indirect N fixation can have a host of unintended consequences from acidification to biodiversity loss to enhanced warming to hypoxia. Clearly, too much N is bad from local to global scales.

The authors set their limit as 35 MT N y-1. In the longform version, they state this value is " ~25% of the total amount of N2 fixed per annum naturally by terrestrial ecosystems" Estimates are that we've exceeded this by a factor of 4.

Basically, they suggest reducing N fixation by 75%.


The authors state that this value is "is a first guess only. Much more research and synthesis of information is required to enable a more informed boundary to be determined."

If you take the negative consequences on N fixation one by one, the key question to ask is what have the trends in these states been over time at the global scale.

Not easy to do.

One of the more tractable is to ask whether N availability to plants, a key component of terrestrial eutrophication and acidification, has actually been increasing or not.

One thing ignored in the paper is that CO2 concentrations in the atmosphere have also been rising. If N2 fixation were drastically reduced, could N availability actually decline as plants receive more C than N?

To answer those, we need better monitoring of N availability to plants at a global scale.

And there is nothing in place to try to answer those questions.

What did bison once eat in November in Kansas?

Most people think the greatest mystery with bison is "How many bison were there before Europeans arrived"?

That number to me seems trivial. If it was 5 million or 20 million is not unimportant, but it's a quantitative question, not a qualitative question.

Today I was out at Konza today. It's November. A warm day, but almost everything is brown. A few rosette forbs were green in the uplands. Some scattered grasses and sedges in other places. Almost nothing good for bison to eat.

Except by the roadsides and some of the firebreaks. Those areas are kept mowed throughout the year and tend to be dominated by the annual grass Bromus arvensis--japanese brome.

The thing about japanese brome is that it greens up early and stays green late. It's an order of magnitude more nutritious than almost anything else out there right now.


The other cool-season grasses just aren't that similar to japanese brome. They are mostly bunch grasses and aren't green or at least vigorous throughout the Kansas winter.

As you can see, the bison really work hard to get it and it must provide an important part of their current diet. As you can hear, they really rip at the turf. The grass there might have been 15 mm tall.

The only thing is that japanese brome didn't use to be in Kansas. When it arrived is uncertain but probably not until the early 1900's.

One thing I've wondered is that if japanese brome wasn't introduced, what would the bison be eating? What would be different about our bison if the grass wasn't there?

Those are just two of the the qualitative questions I think are pretty interesting about bison these days.

Revisiting foliar N isotopes

In 2009, I led a paper on foliar N isotopes. In it, we synthesized >10,000 samples to come up with some global patterns, but also looked at individual studies that related foliar del15N to indices of N availability. In general, we said, N availability and plant del15N scale positively, but there were exceptions. 

In my mind at least, I remember thinking it doesn't always work, but looking at it again, it didn't work sometimes when potential N mineralized was measured. It always seemed to work with actual measurements.

I replotted all the actual measurements of Nmineralization on to a common scale (0-1) and centered the foliar del15N across sites (gave them the same mean of zero).

When you do that, it looks like this:

That's a pretty good summary of almost 4‰ increase in del15N across N availability gradients within sites.

Actually, my first attempt at showing the N availability data was a 1-panel summary rather than 16 panels.

Looking back, reviewers didn't like the 16 symbols and I thought 16 panels were necessary. Reviewers always seem to ask for more data, more panels, more analyses. They rarely tell me (at least) to simplify. 

The 16 panel graph is pretty good. It's clear I should just not have shown the potential N mineralization data in the same graph. 2 out of 3 times they didn't work and should have been shifted out. 

Either way, the 1-panel graph is still a pretty good summary of the data....

Teaching Biogeochemistry and Ecosystem Ecology

I surveyed approximately 150 colleges and universities in the US for how they teach biogeochemistry and ecosystem ecology.

I'll submit the full report to the ESA Bulletin, but can attach a draft of the report here.

Here are a few highlights:

• Less than half of the surveyed institutions teach a course that centers on BiogeoEcosystem science.
• 40% of these were only offered to graduate students. 
• Approximately 1300 students a year take these courses in the US. 40% are at just 10 institutions. 
• 75% of the classes have no lab or field component.

I'll admit that almost half offering these courses is more than I thought. I had it pegged at 25%. Still what is being offered is in line with expectations.

One thing that is clear is that most US institutions do not offer undergraduates access to rich BiogeoEcosystem courses.

Whether the current level of access to BiogeoEcosystem subject material at the national scale is sufficient is a more difficult question. Clearly defined goals and benchmarks need to be defined for that.

One opinion though is that considering the fundamental importance of an understanding of BiogeoEcosystem science, it seems like access can be improved.

• More institutions need to teach more students earlier. 

• On-line courses are needed to supplement the curricula of institutions that cannot provide access to these courses. 

• BiogeoEcosystem students need richer experiences that include lab and field experiences.

There aren't many levers to implement recommendations like these at the national level, though.

Hopefully, information like this will help.

Top ten nitrogen papers

Someone the other day made a statement that some recent work was one of the most important papers on the N cycle published in the last 5 years. Statements like that are hard to refute, but I wondered what that list would like.

I had a short list in my head of papers I could think of. There has been a lot of work on N limitation, new developments in our understanding of microbial processes, and continued advances in the consequences of N deposition.

Most important?

Not sure

I did a search for “nitrogen” and “cycle” and sorted by number of times cited. Then did a search for “nitrogen” in Science, Nature, or PNAS and sorted by # times cited and then relevance. For each I looked at the top 100 papers. I then pulled papers relevant to terrestrial N cycling, but left out reviews and anything with my name on it. I didn’t try to judge 2012 papers.

Here are the 12 papers I came up with for the "Top Ten" list…

Effects of nitrogen deposition

Clark, C. M. and D. Tilman. 2008. Loss of plant species after chronic low-level nitrogen deposition to prairie grasslands. Nature 451:712-715.

Janssens, I. A., W. Dieleman, S. Luyssaert, J. A. Subke, M. Reichstein, R. Ceulemans, P. Ciais, A. J. Dolman, J. Grace, G. Matteucci, et al. 2010. Reduction of forest soil respiration in response to nitrogen deposition. Nature Geoscience 3:315-322.

Mulholland, P. J., A. M. Helton, G. C. Poole, R. O. Hall, S. K. Hamilton, B. J. Peterson, J. L. Tank, L. R. Ashkenas, L. W. Cooper, C. N. Dahm, et al. 2008. Stream denitrification across biomes and its response to anthropogenic nitrate loading. Nature 452:202-U246.

Thomas, R. Q., C. D. Canham, K. C. Weathers, and C. L. Goodale. 2010. Increased tree carbon storage in response to nitrogen deposition in the US. Nature Geoscience 3:13-17.

Global patterns of N cycling

Houlton, B. Z., Y.-P. Wang, P. M. Vitousek, and C. B. Field. 2008. A unifying framework for dinitrogen fixation in the terrestrial biosphere. Nature 454:327-330.

Davidson, E. A. 2009. The contribution of manure and fertilizer nitrogen to atmospheric nitrous oxide since 1860. Nature Geoscience 2:659-662.

Clarisse, L., C. Clerbaux, F. Dentener, D. Hurtmans, and P. F. Coheur. 2009. Global ammonia distribution derived from infrared satellite observations. Nature Geoscience 2:479-483.

Nitrogen limitation to productivity

Elser, J. J., M. E. S. Bracken, E. E. Cleland, D. S. Gruner, W. S. Harpole, H. Hillebrand, J. T. Ngai, E. W. Seabloom, J. B. Shurin, and J. E. Smith. 2007. Global analysis of nitrogen and phosphorus limitation of primary producers in freshwater, marine and terrestrial ecosystems. Ecology Letters 10:1135-1142.

LeBauer, D. S. and K. K. Treseder. 2008. Nitrogen limitation of net primary productivity in terrestrial ecosystems is globally distributed. Ecology 89:371-379.

Soil N cycling advances

Di, H. J., K. C. Cameron, J. P. Shen, C. S. Winefield, M. O'Callaghan, S. Bowatte, and J. Z. He. 2009. Nitrification driven by bacteria and not archaea in nitrogen-rich grassland soils. Nature Geoscience 2:621-624.

Manzoni, S., R. B. Jackson, J. A. Trofymow, and A. Porporato. 2008. The global stoichiometry of litter nitrogen mineralization. Science 321:684-686.

Morford, S. L., B. Z. Houlton, and R. A. Dahlgren. 2011. Increased forest ecosystem carbon and nitrogen storage from nitrogen rich bedrock. Nature 477:78-81.

Summary of these papers:

N deposition causes loss of species even at low levels, it reduces soil CO₂ production, increases stream denitrification, and increases C sequestration in trees.

Nitrogen isn’t fixed in cold places because of enzymatic limitations, nitrous oxide concentrations have been driven by anthropogenic N, and India produces a lot of ammonia.

N and P are limiting everywhere, N is limiting everywhere.

Archaea aren’t important in nitrification, microbes are less efficient with C in high C:N litter, plants can get N from rocks.

Which of these is the most important? Probably not necessary to try and answer.

Did I miss any?

Reciprocal illumination

Apparently, a number of people were more familiar with this phrase than I was...which is to say I had never heard it used before.

In general, it is used to denote the idea that different disciplines can inform one another and help generate advances collectively.

One use is:
http://evolvethought.blogspot.com/2005/02/reciprocal-illumination-of.html

It seems to be a more common term in phylogenetics.

Soil C patterns in the US (and world)

When working with the global soil 15N database, we put together global relationships between soil C and climate as well as texture.

I hadn't seen this put together before, but it turns out they were for the US.

The first paper on this was from Guo et al. 2006 in SSSAJ. The second one was by Homann et al. in 2007 in Biogeochemistry.

I think the Homann paper does a better job of showing relationships than Guo, though neither would necessarily be ready for a text book. The both try to show too much of the dependency rather than the generality.

In short, the authors use data on soils from across the US that had been compiled earlier. Examining the patterns region by region, They then show that clayey soils from wet, cold have higher total C in the top 20 cm (mineral soils) than sandy soils from hot, dry sites.

At some point, someone needs to do a better job of showing this so the figures can get into textbooks.

That said, the basic patterns that we see with our global database match pretty well with the US patterns.

For example, from 5 to 15°C, soil C concentrations declines by roughly 50% in the global dataset, which is pretty similar to what Homann saw for US regions.

For our data, log soil %C declines at a rate of 0.011 per °C. They found a decline of 0.016. Our data showed log soil %C increased 0.49 per log unit of MAP. They showed 0.33. 

Differences can be hashed over, but the global patterns with our somewhat ad hoc data [it was put together for other reasons that soil C] seem to be pretty robust.

Something like what is directly above needs to be in a textbook...

Ecology humor: desert edition

[To be read slowly, dead-pan, with little inflection. Should take about 10 seconds to read. Pause a second with each comma and period.]

About a month ago, I got a cactus. 


A week later, it died. 


I was really depressed because I was like ‘Damn! I am less nurturing than a desert.'

--Demetri Martin

Iconic figure: global climate space

If I had to vote for an iconic figure from work with which I've been associated, it's the global climate space that Andrew Elmore put together. We used this as the basis for a new Whittaker biome diagram earlier, but when it stands alone it's still pretty special. 

What I like about it is that in just a glance you can appreciate:

• The extent of places with ice (or potentially underlain by ice). There's a lot of land with mean annual temperature < 0°C.
• The general wedge shape of precipitation and temperature. 
• The position of the temperate rainforests that sit around 5°C MAT. 
• The occurrence of land that receives > 4 m of rain a year.  

This is a figure that needs to be in every ecology-related textbook. The map of continents just never shows this quite right.

One neat way to use this graph is to overlay data from research to understand the distribution of climates in a given study.

For example, I'll use the climate space diagram as the backdrop for arraying the points where we had samples for the soil 15N synthesis. 

You can see in an instant how well (or not) you've covered a particular portion of the climate envelope.

For this study, we're still a bit short on cold sites and hadn't represented the temperate rainforests well. Hot deserts are undersampled, too.

Something is off...

I've been wrestling over the past few days with trying to understand why soils from the Tibetan Plateau had soil carbon concentrations that were an order of magnitude lower than you expected based on climate. thought it might be the altitude as most of the Plateau is >4000m. Turns out their soils are very sandy with only 3% clay. That explained it.

Then I got a slug of data from beech forests in NE France. Their surface mineral soils averaged -5‰. Soils shouldn't have a del15N less than 0 ‰, no less -5‰. Beech are ectomycorrhizal and their litter is relatively depleted in 15N, but the leaves of these forests were even more enriched than the soil.

Still haven't figured that one out.

In all, when putting together large databases, you have to check the data as it comes in. I've rerun the statistics for this dataset over 100 times. Every time new data comes in, I check to make sure it fits (or doesn't fit) the pattern. You test your mental hypotheses over and over, while error-checking data. Not too different than long-term data needing to be analyzed every year.

Earlier today, Ben Turner sent some data from the dune fields of Haast, New Zealand.

The paper he sent along had pictures that looked like this:

The lat/long he sent looked something like this in google maps:

Doesn't quite look like temperate rainforest. 

[mostly, just rounding error here. The Haast dune system is just north of the town. Kendra and I spent a night there about 10 years ago.]

http://www.doc.govt.nz/parks-and-recreation/tracks-and-walks/west-coast/south-westland/walks-north-of-haast-township/

Landscape level patterns of soil N flux and N isotopes

This might be of interest to just a few, but I've been working hard to understand the degree to which global patterns of soil del15N represent differences in the proportion of gaseous N loss (fractionating loss hypothesis) vs. difference in the degree of decomposition of soil organic matter (microbial processing hypothesis).

There are few papers that try to contextualize losses of N via gaseous pathways vs. leaching, for example.

There also are few papers that look at spatial patterns of gaseous N flux, local or global.

Velthof 2000 did a great gradient study involving 162 flux chambers along 400 m of slope in a grassland.

This is probably the best study that asks whether spatial variation in gaseous N flux patterns soil del15N.

At least over 4 days of measurements, they found a good correlation between the two.

Best correlations were with shallow soils (0-5 cm). r2 was about 0.4. But correlations for del 15N of soils even just 10-15cm deep were non-significant (r2 = 0.03, P > 0.3)

For what it's worth, they said that "Both the del15N values and N2O fluxes were highest at the steepest part of the slope."

Would steep slopes also have more processed OM?

Velthof, G. L., J. W. van Groenigen, G. Gebauer, S. Pietrzak, S. C. Jarvis, M. Pinto, W. Corre, and O. Oenema. 2000. Temporal stability of spatial patterns of nitrous oxide fluxes from sloping grassland. Journal of Environmental Quality 29:1397-1407.

Global patterns of soil C concentration--an index for relative decomposition rates

Patterns of soil C concentrations as a function of mean annual temperature and precipitation in surface soils for 600+ soils from around the world.

One of the keys to understanding global N cycling patterns is understanding the global patterns of rates of decomposition. There are a few syntheses of litter bag studies, but the net result of inputs and outputs of C for soils are hard to interpret.

One of the best indices (I think) of the rate of decomposition relative to inputs is just soil concentration.

Most syntheses to date have examined soil C content, not concentration. Mostly because the goal is to determine soil C storage. Concentration only helps to determine content.

The major syntheses I can think of--Mac Posts 1982 and Batjes 1996--examine patterns for different soil orders. temperature or precipitation is never on the x-axis.

Bearing that, there should be an advance possible by putting climate on the x-axis and %C on the y-axis.

It seems like this would have been done, but I can't find a graph like it and people that are likely to know about it seems surprised when I show them.

For the 570 mineral soils I looked at, MAT and MAP explain about 60% of the variation in log-transformed soil C concentrations. That's a high r2 given all the variability out there in the world and other factors like how much clay is in soil, the quality of the plants,  and whether plants are eaten or not.

The interpretation of the patterns essentially is that if soil %C is an index of the amount of decomposition of plant biomass relative to primary productivity, then hot, dry places have soils with highly processed C.

The patterns of relative decomposition now seem pretty clear.

The question now becomes whether those hot, dry places consistently are elevated in 15N, such that soil 15N patterns are being caused by relative decomposition rates.

Two tests here.

1) Soil %C and soil 15N should scale across sites.
2) If soil 15N increases with increasing MAT and decreasing MAP, the relationship should disappear after accounting for variation in soil %C.

Global N cycling: the Climate-Nutrient hypothesis

Patterns of soil 15N and P availability in the Amazon. From Quesada et al. 2010.

When we looked earlier, the degree of decomposition affects soil organic matter content and the isotopic ratio of the N in the soil.

For individual classes of organic matter, be in leaves, organic layers, mineral soils, or fractions of mineral soils, the more microbes process organic matter, the more C is lost and the more enriched the nitrogen becomes in 15N.

The "processing hypothesis" is a standard explanation for vertical profiles of organic matter in soils. Deep soils have lower C concentrations and higher del15N than shallow soils because the stuff at the bottom has been worked over by microbes**.

**Other mechanisms affect vertical profiles, too. Plants preferentially cycle light N up to the top. Illuviation can also transport C and N downwards.

What applies vertically, could also apply horizontally.

Yet geographic patterns have largely been explained with the fractionating loss hypothesis. Soils enriched in 15N are thought to be enriched because they have lost a larger proportion of their N to fractionating pathways compared to relatively depleted soils.

Two sets of observations come together to generate the main latitudinal patterns.

1) Tropical soils are enriched in 15N compared to temperate soils
2) Tropical soils have high rates of N2O flux.

Put together, the two reinforce one another to solidify a view of latitudinal gradients.

But why would that be?

Nitrification or denitrification are not thought to be temperature sensitive like nitrogen fixation.

Therefore, it's indirect controls.

One of the major hypotheses is the Climate-Nutrient hypothesis. Tropical systems are thought to be more P-limited, which increases the degree of N surplus. Greater N availability increases the likelihood of gaseous N loss.

Quesada's work (above) is a good example of data that calls this hypothesis into question.

Within the Amazon (which is all hot), across a gradient of P availability, 15N is lowest in low P soils, not high P soils. Low P soils are supposed to have the greatest excess N and the most gaseous N loss.

To maintain the Climate-Nutrient hypotheses, explanations get pretty complicated. Rates of N2-fixation by plants have to vary in ways that one wouldn't expect. Or losses have to become episodic and almost catastrophic. 

A number of other questions come up. Recent work suggests that N losses via NO3- leaching or dissolved organic N loss to streams in  tropical systems can be high, too. And N2O is just one of the gaseous fluxes of N. Denitrification also produces N2, which is nearly impossible to measure.

Are tropical systems losing a greater proportion of their N via gaseous pathways? That part has never been quantified directly.

There is enough evidence out there to at least question the traditional view of the fractionating loss hypothesis driving global patterns in soil 15N, if not our views of the N cycle in the hot, cold, wet, and dry.

The next question is whether the processing hypothesis can explain more variation with fewer mechanisms. 

If so, global patterns of N cycling need to be reconsidered.

Solving the riddle of the Amazon's nitrogen cycle

Sand vs. clay content (averaged 0.1° latitude/longitude) for the global soil 15N data set (so far). Red points are the Amazon. The one high blue point is from Panama.

When examining all the soil nitrogen isotope data from around the world, the Amazon rainforests always stood out as being too enriched based on everything else we knew about it. The Amazon is hotter and wetter than most other places, but it's del15N of soil organic matter was still 3‰ too high.

Something was different about the Amazon rain forests. 

There was either more gaseous N loss coming out of there than expected or more microbial processing that was enriching the soils. 

But what was it?

I looked at soil pH. Nothing unique there.

Soil carbon and nitrogen. Still couldn't explain the Amazon effect away.

One "outlier" for the Amazon has always been the white sands forests. Their 15N was never as enriched and those forests were always considered "nutrient poor".

Here's what Quesada et al. 2010 recently wrote on the topic: 

"That nitrogen is in excess for most tropical ecosystems (except forests growing on white sands) is, of course, already widely accepted, but from our pedogenic viewpoint, however, we also argue that as these losses continue to occur, the soil C:N ratio of the soils should also gradually increase (Fig. 12c) until a critical threshold around 30 mg kg−1 is reached. Beyond this, the system once again becomes more closed with respect to the nitrogen cycle, and thus enrichment ceases and the soil _15N tends back towards the isotopic composition of the input precipitation (ca. 0‰). The white sands forests of Amazonia have long been known to have relatively depleted _15N (Martinelli et al., 1999), and a large fraction of the sites on the “downside” of the curve are indeed sandy soils. But importantly, there are also other soils types with similar low P levels on the low side of the “breakpoint” (Fig. 11). This suggests that rather than just being a soil type effect, extremely low soil P concentrations may itself be the cause of these soils having strongly depleted _15N, with the phosphorus shortage itself leading to a lower rate of nitrogen return to soil. That is to say (in simplistic terms) at very low P availability, the forests actually become nitrogen limited."

When I look at the data, the one thing that stands out about the Amazon is the high clay content.

There just are few places in the world with such high clay concentrations. 

The white sands of the eastern Amazon might be the outlier for the Amazon, but it's the "black clays" of the western Amazon that are the outlier globally.

Quesada's argument about C:N and P might have some validity (I can work through that later), but the once you take into account clay concentrations, the Amazon effect almost disappears (<1 average="average" enrichment="enrichment">

1‰ is tolerable.

So how clay impacts soil 15N will be important. Does it control P availability and decomposition patterns that impact gaseous N loss? 

Or is it by impacting retention of enriched substrates and microbial processing?

Answering this question will get at whether Amazonian rainforests have a unique N cycle or not, but we at least now a proximal driver.

Beginning to reinterpret global soil N cycling patterns

If there is a state-change in how we think about global N cycling, or at least interpret soil 15N, it's likely to come by looking at covariates. For example, it's been hypothesized that hot sites have higher 15N because they have lower soil P, or higher soil pH, than cold sites, which is more directly driving large-scale patterns.

Or it's something else we haven't considered yet.

Working through the older literature on soil 15N, one paper that stands out is Knute Nadelhoffer and Bryan Fry's 1988 paper. There, they sampled a long-term experiment that varied litter inputs to soils at the University of Madison Arboretum. They took the soils and incubated them for 600 d, periodically leaching the soils.

Their main graph shows two things happen during incubation as microbes processed the soil. N concentrations declined and 15N went up.

I regraphed their data using %C instead of %N and you see something similar. Over time, C is being respired, C concentrations drop, and 15N of the SOM goes up.



About 80% of the N that was lost was from inorganic N in leachate. The other 20% was presumably from gaseous N loss.

The general idea that microbial processing of organic matter enriching 15N of the remaining material and causing C and N concentrations to decline has been shown repeatedly. Litter bag studies (Connin et al. 2001) show the same pattern. So do studies of different fractions of soil organic matter that differ in their degree of microbial processing (Kramer et al 2003):

The microbial processing concept has been applied to understanding vertical patterns in soil 15N, but never geographic patterns. 

It's possible that hot, dry sites are not (potentially) enriched in 15N because they lose a greater fraction of N to gaseous N loss, but instead because their soil organic matter is more decomposed.

To test for this, the key would be to look at soil C or N concentrations. C:N is generally considered an index of decomposition, but Nadelhoffer and Fry showed that as decomposition proceeded, both C and N declined. C concentration (or maybe N concentration) might be a better (not perfect) index of microbial processing. 

If the processing hypothesis predominates, then hot, dry sites would have lower %C and higher 15N. After accounting for variations in %C, if mean annual temperature or precipitation no longer predict soil 15N, this might be a good indication that they impact the degree of organic matter processing more than the relative importance of fractionating losses.

Once you get to the point of competing hypotheses, the only thing left to do is test them. 

Global soil 15N: the most recent synthesis was 10 years ago

Soil nitrogen isotopes have the potential to elucidate long-term patterns of N cycling. In short, the ratio of 15N/14N in soil organic matter should differentiate soils that lose more N through fractionating pathways like gaseous N loss vs. those that primarily lose N through non-fractionating pathways such as organic N loss or NO3- leaching.

In 1999, there were two major syntheses that were at odds. Handley's synthesis stated that high precipitation sites had low soil 15N--except it wasn't significant. They did show good relationships with latitude. Martinelli showed that tropical forests differed in soil 15N from temperate forests, which again highlights the importance of latitude.

The only problem is that the two papers showed different results. Handley had low-latitude soils with lower 15N, Martinelli higher.

Taken at face value, relative importance of denitrification should be higher in the temperate systems if one follows the Handley data, tropical systems for Martinelli.

A few years later, in 2003, Amundson et al. conducted a new synthesis. Like Handley, the data were global and across all ecosystem types. They examined soil 15N to 10 cm and 50 cm. They also calculated "regional" averages, which diminished the potential "pseudo replication".

The synthesis regressions, shown above seemed to have solved the confusion. Latitude wasn't the driver of soil 15N, it was climate. Hot sites had high soil 15N. So did dry sites.

From this, they concluded,

"Because most undisturbed soils are near N steady state, the observations suggest that an increasing fraction of ecosystem N losses are 15N-depleted forms (NO3, N2O, etc.) with decreasing MAP and increasing MAT. Wetter and colder ecosystems appear to be more efficient in conserving and recycling

mineral N."

The paper seemed to calm the waters on patterns and interpretations. Handley's earlier results on precipitation were supported. The latitude question was resolved. Interpretations were based on careful modeling of isotopic dynamics.

Or not. 

A few problems.

First, there were still no data that supported their interpretations that low soil 15N was associated with "efficient" conservation and recycling--the "openness" argument from earlier. They should have said that cold, wet systems lose a smaller fraction of N through fractionating pathways like gaseous N loss. The N cycle could still be "open" and lose a lot of N relative to rates of cycling, just lose it through DON or NO3- for example. 

Second, they still hadn't shown that precipitation had significantly impacted soil 15N. The P value for the 10-cm soil 15N samples was 0.14. I reanalyzed their data here:

For 50 cm, the P value was 0.09. Still hardly definitive.

For global syntheses, the number of soils examined was still incredibly small. For 0-10 cm, just 85 sites. For 0-50 cm, just 47. Remember, in 1978, Shearer et al. had done over 100 soils from the US alone. 

The lack of replication isn't a personal fault of the authors--although they could have asked Georgia and Danny for their raw data**. As a discipline, although there had been major questions about global patterns of N cycling, there was never a global effort to nail them down.

**Apparently no one ever asked them for their data. Danny recently wrote me "Although I was not certain, it turns out that I disposed of all of the raw data when my office was moved recently.  You, of course, have access to the paper. I'm afraid that is all there is. I never thought for one instance that anyone would ever be interested in those original data. I am sorry that I cannot be more helpful and sorry that our tediously acquired data will not be part of your data base."

The reason that the low replication is important is that the results are still highly sensitive to a few points. If I exclude 3 points from the regression, MAT is not significant.

So, where does that leave us? 

1) No one has yet to show that high precipitation sites would differ inherently in long-term N cycling characteristics than low-precipitation sites.

2) Hot sites have been shown to have significantly higher soil 15N than cold sites, but it's tenuous.

3) The interpretations of soil 15N are still shaky. Even accepting the conventional wisdom on how to interpret soil 15N, what these patterns implicate about the N cycle have not yet been resolved.

Lastly, there is one more thing that is conspicuously absent from all of these syntheses:

Carbon. 

The C and N cycles are tightly linked, but the syntheses have been done in absence of understanding carbon.

In a bit, I'll show that also looking at carbon has the potential to fundamentally change our interpretations of global N cycling.