Reading the frontier: A year in our journals

An enjoyable consequence of writing “President’s Picks” series featuring SSSA research this year is that I regained the habit of consistently reading through all of our journals’ tables of contents and their abstracts just for fun. This allowed me to learn things I might not have found in my previous years’ normal routine. I started the series as a way to celebrate research in our journals, especially work that warranted rereading and changed a perspective I previously held. By late summer, I had accumulated enough research picks that I started to reread them back-to-back rather than as separate, isolated monthly choices. The papers range from molecular interactions on mineral surfaces to 15-m-deep vadose zones and Martian regolith simulants. On the surface and spaced out in time, one might not expect them to have much in common. However, several ideas appeared repeatedly, which inspired me to dedicate this month’s column to them. 

History, sequence, and state

Not surprisingly, the clearest pattern was in studies where a soil process or observation depended strongly on what that soil recently experienced—its own history. Li et al. (2026) showed that antecedent rainfall and repeated freeze–thaw cycles had opposite effects on subsequent wind erosion in an Aeolian sandy soil. Rainfall strengthened the soil surface and formed a protective crust, whereas repeated freeze–thaw weakened the soil and progressively increased wind erosion through the tested cycles. O’Donnell et al. (2026) demonstrated the importance of prior conditions over a much longer period while reconstructing more than 80 years of sediment accumulation in an agricultural wetland. Their sediment record reflected interacting land-use transitions and episodic climate events rather than simple, gradual accumulation through time.

The August and September Picks examined related effects in deep soil and biogeochemical processes. Sedaghatdoost and Mohanty (2026) sampled as deep as 15 m in the Gulf Coastal Plains and found a deep vadose zone that was much more biogeochemically active than the common assumption of conservative transport below the root zone would suggest. Wetting and drying altered iron redox chemistry and, through its coupling with phosphate sorption, allowed the deep soil to shift from a phosphorus sink to a phosphorus source depending on recent hydrological conditions. Huang et al. (2026) likewise found that the frequency of wetting and drying changed microbial processing of dissolved organic matter in lake-wetland soils, including aromatic compounds usually considered relatively resistant to degradation. The authors proposed that repeated cycling disrupted aggregates and exposed carbon that had previously been physically protected.

Chen et al. (2026) showed another form of history dependence while examining microbial clogging under variably saturated flow relevant to managed aquifer recharge. The hydraulic response to microbial clogging did not decline monotonically toward one fixed conductivity, but passed through a rapid decline, a partial rebound and readjustment, and eventually a long-term dynamic equilibrium. Moderately unsaturated conditions also promoted more microbial growth and extracellular polymeric substance (EPS) accumulation than fully saturated conditions, while the clogging itself was vertically stratified, with bacterial cells dominating the aerobic near-surface zone and EPS occluding micropores below. Cramer et al. (2026) used neutron and X-ray imaging to document a related change around microplastics, where water initially moved more slowly through water-repellent zones and later bypassed them through preferential pathways. In both studies, the measured response developed over time and could not be explained solely from the initial condition.

Two other Picks considered related issues through modeling approaches. Mair et al. (2025) modeled root water uptake and showed that the influence of rhizodeposits depended on the intensity, timing, and frequency of atmospheric boundary conditions. Liu et al. (2026) examined fingering flow, where infiltrating water can self-organize into fast, narrow pathways even in a uniform porous medium without pre-existing macropores, and an optimality-based active region model in which hydraulic conductivity depends on flux rather than saturation alone. Both papers are great demonstrations of how fixed soil properties alone are incomplete and insufficient to explain the response. 

We have always known that soils change, but I think these papers do a great job of describing how sequence, frequency, and prior state are essential in what governs the outcomes of soil processes. In other words, this means a measurement at a single time may not contain all the information needed to understand a subsequent response. 

The measure is not the mechanism

My March Pick, by Widanagamage and Patrignani (2026), compared laboratory- and field-determined soil water retention curves in a well-aggregated tallgrass prairie soil. The field measurements did not necessarily provide a better representation of the wet end of the curve, where sensor measurement frequency and equilibration lag affected what could be measured under strong aggregation and nonequilibrium flow. I particularly like that result because it demonstrates and reminds us that each measurement is produced using an instrument, a protocol, a sampling interval, and a scale. Those choices affect the observation and the context of our interpretations whether we acknowledge them explicitly or not.

Kogler and Nordt (2026) examined a related issue at a different scale. Along a Texas Vertisol climosequence, gilgai microhighs and microlows separated by only a few meters differed systematically in their organic and inorganic carbon stocks, and that meter-scale variability obscured the broader relationship with precipitation that might otherwise appear at regional scales. Freiberg and Arai (2026) examined the issue at molecular scales by showing that the configuration of functional groups in six carboxylic acids, rather than concentration alone, strongly influenced how effectively the acids dissolved hydroxyapatite. Ayetan et al. (2026) found a comparable depth-dependent response in soil compaction, where lower tire inflation pressure substantially reduced stress in the upper profile while stresses at depth remained strongly influenced by total wheel load and soil structural condition. Kiałka et al. (2026) showed a similar issue when scaling soil processes into global models. The effects of soil macroporosity were suppressed when precipitation and hydraulic properties were averaged across large model grid cells but became much more apparent when subgrid variability was represented. In each case, a number, setting, or measurement provided useful information, but did not by itself describe the full reality of what governed the system. 

I probably noticed this theme among the Picks because my own group has spent a fair amount of time examining similar issues. This year, we published our perspectives on distinguishing comparative soil health benchmarking from mechanistic prediction, how soil health is a nonequilibrium flow problem, and how limitations arise when emerging contaminants are studied as single species even though environmental exposures occur as complex mixtures (Daigh et al., 2026a, 2026b; Daigh, 2026a). We also published a paper reconsidering how we define soil and which process-based states should be included in that definition (Daigh, 2026b). Obviously, those papers were not part of my President’s Picks, but they probably help give some insight or explain why I selected several of the Picks discussed here. 

There is nothing inherently wrong with reducing a complicated system to a useful number for the strict purpose of benchmarking. We use field capacity, wilting point, carbon stocks, soil health scores, weathering indices, and many other abstractions because they let us compare, communicate, and make efficient decisions. Problems arise when those abstractions are applied as a predictive means outside the conditions and context for which its relationship to the governing process has been established. The papers above examine that issue at different scales, but each helps remind us to ask what information a number represents and what information it does not. That seems especially important now as more of our measurements are aggregated into large datasets, remote products, decision-support tools, and models used beyond the context where the measurements were made.

Defining the scope of soil science

The April and October Picks addressed research involving Martian and lunar regolith simulants. Coker et al. (2026) evaluated thermal treatment, aqueous leaching, and biological reduction for removing perchlorate from a Martian regolith simulant, a remediation problem directly relevant to potential extraterrestrial crop production. Fackrell et al. (2026) showed that commonly used lunar and Martian regolith simulants contained Earth-derived microorganisms and that some sterilization methods altered the material. Their furnace treatments, for example, collapsed smectite and dehydrated gypsum in simulants containing hydrated phases, while packed autoclave bins could retain viable organisms in their centers where compaction limited steam penetration. If preparation or sterilization alters the analog material, that effect becomes part of the experimental design and interpretation.

Palma et al. (2026) examined a terrestrial environment that may serve as an analog for Martian regolith. Their work on James Ross Island in Antarctica described cold, arid, salt-rich soils that are weakly weathered, including evidence of possible dry permafrost and complications in applying the Chemical Index of Alteration where sodium enrichment depresses the index. I find studies like this fascinating because they test familiar pedological methods under environmental conditions that differ markedly from those in which those methods are most often applied. In some cases, the methods remain useful, while in others, the environmental conditions indicate limitations or require modified interpretation. Along a related frontier, Morris et al. (2026) approached the boundaries of soil science from the microscopic direction, reviewing how fungi physically and chemically interact with minerals and how difficult it remains to connect those interactions across scales. They also considered how fungal weathering features and biominerals could serve as potential biosignatures when interpreting extraterrestrial materials.

Patterns across this year’s President’s Picks

Taken together, these Picks do not point to a single new theory or coordinated research agenda. Instead, they show investigators working on very different problems while testing assumptions that are often simplified in standard descriptions. A deep vadose zone exhibits substantial biogeochemical activity, antecedent wetting and drying affect subsequent responses, meter-scale microtopography can obscure regional carbon relationships, and regolith simulants can contain terrestrial microorganisms and be mineralogically altered during sterilization. In each case, examining the system in greater detail identified conditions under which a common assumption or simplified representation was incomplete.

This relates to something I wrote about in my July president’s column regarding the depth problem in soil science. Our discipline is at its best when we accumulate cautious and consistent evidence across many soils, sites, seasons, scales, and decades. That caution should include periodically testing the quantities, boundary conditions, indices, and assumptions that have become standard practice. This applies to students and early-career researchers, but also to those of us who have worked in the field long enough to have developed our own familiar approaches. A parameter that is routinely fixed, an index that is widely accepted, or a spatial scale that has become standard may be entirely appropriate… BUT, it is still worth knowing the conditions under which it remains appropriate.