Path Dependencies Between the Atmosphere and the Economy:
Seasonal to Subseasonal Signals for Global Food Security
If agriculture fails, everything else will fail.
This statement is attributed to the late Dr. M.S. Swaminathan, one of the world’s leaders in the green revolution and colleague of the late Dr. Norman Borlaug. I would slightly modify the statement by adding “energy and water,” but this cannot be underestimated as a fundamental pillar of a stable society. We might think that the winners of the ‘AI wars’ will be those countries that possess technical superiority with respect to computational and data sciences and capabilities. But without access to a stable and cost effective agriculture and energy production and distribution system, technological wins will only be temporary.
Natural resources are, and will remain, the cornerstone of the global economic, socioeconomic, and geopolitical engine — with particular emphasis on food.
While artificial intelligence and large language models continue to dominate short-term market behavior — capturing the aspirations and imagination of investors, entrepreneurs, and policymakers alike — there exists a series of more fundamental risks and opportunities that, while receiving less attention, are far more consequential when viewed through the lens of economic and geopolitical security. As the threats and opportunities shaping the global agricultural system have become more apparent in recent years, so too has the recognition that preparing for new and emerging supply-side risks can offer a competitive advantage to those ready for impending volatility.
Within the last month, there have been disruptions to food supply and/or price in dairy, poultry, grains, sweeteners, and energy, whose ripple effects are not contained to the country of origin.
In recent years, risks have emerged which have been shown to stress both production and trans/intercontinental distribution of food, fiber and material commodity derivatives. Supply side risks are triggered by any number of catalysts; among them, weather/climate disruption, geopolitical activity, and natural/manmade disasters. Predicting these disruptions is a fool’s errand. Preparing for them, however, can be immensely valuable.
The challenge in risk identification used to be a lack of data. Fast forward a couple of decades, and the risk is now one of abundance. The sheer volume of real-time information available to decision-makers can often obscure more than it reveals.
False positives and overfit models are ubiquitous. For individuals and organizations tasked with managing physical and financial agricultural risk, traditional tools and techniques are no longer sufficient.
Understanding today’s increasingly complex global supply chains requires a new ontology — one that enables the clear separation of signal from noise.
Research into the risks and opportunities associated with the global agricultural value chain shows that today’s physical commodity analysis is increasingly incorporating alternative data and techniques to support decision-making. However, many of these tools remain outside the reach of those directly responsible for risk management.
It is time to adopt a new framework for understanding and managing agricultural risk — one that combines discretionary, scientific, and quantitative approaches. This fusion offers a more distinctive perspective on the where, when, and why of global commodity dynamics.
As food and natural resource security rise on the priority list for both nations and corporations, such an evolving framework should be expanded and adopted across commercial and geographic sectors.
While the focus here is on risk applications, the same methodology can also be used to identify regions where threats to agricultural productivity and crop yields may lead to food insecurity among vulnerable populations.
First Principles Data
One way to evaluate agricultural commodity production potential is through a mass balance approach that leverages physical climatological teleconnections. Energy transfer in one region often manifests in another — across varying scales of lead and lag time between cause and effect. To better understand these complex dynamics, we need to break them down to first principles. The climate-energy system remains one of the most intricate puzzles we face.
Analysts already apply this type of thinking at a high level, but there’s significant room for improvement with the right mix of data, science, and methodology — particularly when integrating physical, biophysical, and economic variables. For example, when major meteorological agencies signal an incoming El Niño event, agricultural traders often begin positioning for anticipated impacts: drier-than-normal conditions in Southeast Asia, warmer trends across North America, and corresponding shifts in crop development and yield.
Of course, since the climate system is inherently chaotic, these outcomes don’t always play out as expected. Still, general patterns tend to follow phase shifts, offering just enough signal for those equipped with the tools and foresight to act early.
Taking this logic further, we can blend empirical and reanalysis-based physical time series data with pattern recognition models to construct forward-looking scenarios for supply, distribution, and price. This allows for a more nuanced understanding of the physical and operational mechanisms driving agricultural productivity and yield — by commodity and variety — rather than relying solely on often fallible econometric models.
There are also well-established relationships between macroeconomic variables and global commodity prices. For instance, a strong inverse correlation typically exists between USD strength and global food prices (as illustrated by the USD Index and the Global Food Price Index from 1990 to 2018).
However, while these macro relationships are essential, they don’t tell the full story. Many traditional risk models already account for such variables. In that sense, the market becomes the scoreboard — but understanding what drives the plays still requires a deeper, more integrative approach.
After using this generalization as a starting point, we can extrapolate these types of rules to construct a framework that incorporates specific geospatially referenced physical and biophysical variables, many of which are not necessarily within the purview of most market participants.
In order to gain an edge in identifying potential gaps in physical and biophysical expectations, we turn to non-traditional data sources, which, by and large, are not exploited via both traditional discretionary and quantitative techniques. In this case, we suggest a move to incorporate global climate indices.
Additional publications will go into greater detail emphasizing (a) the incorporation of additional physical and satellite derived time series data as well as macroeconomic data, and (b) more granular relationships per commodity category.
Primary Climate Indices — Steering Currents
Prudent strategies that target supply chain risk should be predicated upon identifying climate variables that serve as enablers or limiting factors, with respect to agricultural productivity (total mass) and yield (extractable quantity per area unit). One approach that is part of this framework attempts to utilize non-standard climate index data, typically collected and maintained by one of several global governmental meteorological agencies. As we focus our approach on time series global climate indices, we attempt to identify precursors that serve as catalysts to forward price behavior, providing an advantage over conventional analytical methods.
Proprietary research has uncovered strong relationships which connect the physical variables that are most closely correlated to global agricultural productivity across all commodity asset classes. These relationships also obey the laws of physics, so they are scientifically defensible. This is not always the case with quantitative models.
The remainder of this discussion below summarizes the climate variables which have been selected, and their relationship to crop growth; a summary table for quick reference is included. We then describe in more detail two oscillations that are key drivers for agricultural productivity and food security: the El Nino-Southern Oscillation, and the Quasi-biennial Oscillation. Future research will be expanded to include additional physical climate indices, and also other forms of data such as satellite imagery, field trial data, and sensor data, in support of the construction of a more granular global database of causal mechanisms and their descriptions.
These relationships will serve as the ‘rules’ for machine learning implementation and scenario planning.
El Nino-Southern Oscillation (ENSO) and the Quasi-biennial Oscillation (QBO)
In addition to macro variables which are partial price determinants, the time series index data in the table below was selected via discretionary research supported with peer review literature, to have the most significant impact on agricultural yields, supply, and price.
ENSO variables are among the primary selective pressures influencing the variation and distribution of temperature, moisture, and humidity across the world’s key agricultural regions. When examining ENSO phase, we begin with the equatorial Pacific Ocean, which acts as the central reference point for assessing both the phase and its transition.
An El Niño pattern refers to a warm phase of Pacific Ocean sea surface temperatures (SSTs), while the opposing La Niña phase is marked by cooler Pacific SSTs and a typically warmer Atlantic Basin. Each phase has a distinct influence on the jet stream, which plays a crucial role in controlling moisture availability across much of North and South America.
The temporal behavior of El Niño events can vary considerably, often depending on the interaction between surface and subsurface Pacific SSTs, and the pressure differential between Darwin (Australia) and Tahiti — a key atmospheric teleconnection used to define ENSO strength and development.
When a positive ENSO phase develops, the Nino4 (N4) region in the central Pacific begins to exhibit abnormally elevated SSTs. This warming is typically accompanied by a shift in equatorial tradewinds. Under neutral conditions, these tradewinds blow from east to west (easterly), steering moisture and ocean currents westward across the Pacific.
However, during El Niño, the winds shift west-to-east (westerly), originating near Southeast Asia and driving convection and moisture toward the Americas. This displacement of moisture away from the Asia-Pacific region often results in significant deficits for agricultural zones across Southeast Asia, Australia, and surrounding regions.
The outcome is an increased risk of moisture stress and yield loss — factors that can cascade into regional food insecurity and supply shocks. Understanding these mechanisms is critical for anticipating agricultural impacts and preparing proactive responses.
As stronger El Nino patterns develop, the warm surface and sub-surface masses influence the amount and distribution of rainfall that is captured by the jet stream and delivered to agricultural belts in North and South America. The volume of warm water which makes its way to the surface increases, and excess evaporative moisture is then captured through convection; the beneficiaries of this additional precipitation (typically Brazil, Argentina, portions of the US and Mexico) receive more rainfall than climatological normal. When the opposite La Nina pattern is dominant, these regions experience moisture deficiencies. El Nino phases tend to operate on cycles anywhere between 8 and 24 months; in extended El Nino events, we often see two positive phases within the cycle.
The phase of QBO is also important to global agriculture interests, as this serves to dictate where extended regions of high and low pressure will situate. High pressure regions serve to block moisture from entering an area; conversely low pressure anomalies are oftentimes associated with enhanced storm tracks and above normal moisture. Transitions between west and east phase QBO generally occur of a frequency of 7–12 months.
Periodicity has been observed to become more active both in terms of phase and amplitude. As with ENSO, there are specific QBO signatures associated with moisture delivery to agricultural origins and a better understanding of the loading patterns behind these transitions will allow for early identification of target precursors, which impact global agricultural yields.
Summary
Managing risk around global agricultural productivity can be problematic for decision makers in commercial and food security circles. Weather and climate disruption across the agricultural supply chain makes supply-side planning difficult for physical and financial forecasting. As discussed, there are even more important socioeconomic implications to consider when climate phase changes materialize.
Coupling physical, biophysical, and economic data with pattern recognition techniques is a way to anticipate and manage risk in today’s increasingly complex world. Through the incorporation of differentiated Earth observation data and processes, learnings can be mapped to a variety of global agricultural commodity exposures.
As a result, the astute analyst can identify the most relevant combination of macro and climate variables with sufficient lead time to prepare for and proactively respond to emerging threats. The identification of climate precursors to global agricultural productivity is important for developing an early read on supply and demand drivers — and for anticipating inflection points in supply that lead to price reactions ahead of the market.
Numerous physical variables have been identified as key pieces of the signal identification and optimization puzzle. Continued work in this area will provide risk managers with a suite of favorable and geographically specific mitigation strategies.
This approach also minimizes risk and protects against knowledge decay. As new data becomes available daily, weekly, and monthly, new relationships can be discovered and verified.
Future articles will examine many of these topics in greater detail, generating ideas for new strategies to manage risk.
As always, preparation trumps prediction.
— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —
This paper will be published as part of the Path Dependencies Between the Atmosphere and the Economy series.
Dr. Michael Ferrari is the Chief Scientific and Chief Investment Officer at AlphaGeo, founder and Managing Partner at Atlas Research Innovations, and advisor and partner at Weather Trends International.
