Using multielement isotopic analysis to decipher drought impacts and adaptive management in ancient agricultural systems

It has recently been proposed that drought events over the last several millennia affected agricultural systems and triggered the development of irrigation technologies in ancient Near East settlements. This conclusion is supported by carbon isotope (δ¹³C) measurements presented in a detailed study by Riehl et al. (1), who analyze barley grains from 33 archaeological sites, spanning 8,000–2,500 y ago, alongside data from 13 contemporary locations. The study connects plant physiological responses with paleoclimate data to gain valuable insight into early agricultural practices. Although we agree with the authors’ conclusion that drought stress was an issue in many ancient agricultural settlements, we wish to note that other factors could have contributed to the observed differences along with climatic effects on δ¹³C signals.

In contemporary systems, site-specific conditions such as soil fertility and relative humidity are known to cause significant variation in ∆¹³C (δ¹³C normalized for changes in atmospheric CO₂ concentration) (2). For barley, it has been shown that a switch in the predominant form of soil nitrogen causes changes in leaf ∆¹³C as large as 2.6‰ (3). This is critical because, regardless of the large intrasite variation, Riehl et al. (1) use a 1‰ ∆¹³C decline to distinguish stressed from well-watered plants, without accounting for soil fertility effects. Manure has been used to fertilize crops since at least 8,000 y ago, and this is evidenced by a distinct relationship between nitrogen isotope ratios (δ¹⁵N) and ∆¹³C of ancient barley seeds (4). Crop δ¹⁵N can be distinguished from forage values, indicating manure application, accounting for N-source–derived ∆¹³C variation.

Fig. 1 shows that ∆¹³C alone cannot distinguish between variation in nitrogen and water availability. Water status can be affected either by changes in photosynthetic rate or water availability (5). In the absence of actual yields, which could attest for net differences in photosynthesis, a more intricate model including δ¹⁵N and oxygen isotopes (δ¹⁸O) would be required to separate these confounding effects. This is possible because δ¹⁸O is controlled entirely by fractionation of leaf water, regardless of photosynthetic responses (2), and δ¹⁵N by N source (3), which influences ∆¹³C by changing photosynthetic rates. With this information, classic conceptual models (2, 5) can be used to decipher drought impacts and adaptation to climate change while accounting for site-specific environmental conditions.

Fig. 1.

Isotopic methods that constrain the interpretation of variation in ∆¹³C. (A) Greenhouse-grown barley leaf values by Lopes et al. (3). Water treatments are grown with the same nitrogen source. Nitrogen treatments are grown with the same water inputs. Relative shifts in δ¹⁸O are indicated with an arrow for water treatments, whereas no change should occur for varied nitrogen inputs (B) Adapted from Bogaard et al. (4), data from pea (green), barley (red), and wheat (blue) grains. High, medium, and low refer to relative manure inputs, boundaries are general, but established with modern experiments. Importantly, in each experiment, variation in ∆¹³C values caused by varying N inputs is larger than that of water stress and the 1‰ variation used by Riehl et al. (1) to distinguish drought stress and well-watered plants.

Riehl et al. (1) do mention climatic validation using δ¹⁸O from spatially coarse cave and sediment data, yet their study attempts to discern regional drought impacts from localized microclimate effects. By using δ¹⁵N and δ¹⁸O data from the same grains used for δ¹³C analysis, a more refined characterization would be achieved. We suspect that across-site variations in ∆¹³C could be explained by a combination of irrigation and fertilization effects, but only an evaluation of δ¹⁸O and δ¹⁵N would unequivocally demonstrate that this is the case. Our proposed isotopic analysis would also allow for direct comparisons between sites, yielding a more nuanced interpretation of climatic effects on the process of crop domestication and emergence of human civilization.