Decadal climate-driven decoupling between gross primary productivity and tree growth in Mediterranean forests

Using a process-based model across southern Italian Mediterranean forests, this study reveals that decadal climate variability has driven a decoupling between photosynthetic activity and tree growth, where canopy greening can mask underlying structural vulnerability and previsual decline despite apparent trends in photosynthesis.

The Gist

The Big Picture: The "Green Mask"

Imagine a forest as a busy factory.

• The Photosynthesis (GPP) is the factory's assembly line. It's where the trees take sunlight and CO2 to make energy (sugar).
• The Tree Growth (BAI) is the shipping department. It's where the factory takes that energy and uses it to build new wood, making the tree thicker and taller.

For a long time, scientists assumed that if the assembly line was humming along happily (lots of green leaves), the shipping department must also be busy (trees are getting bigger). They thought the two were perfectly linked, like a hand in a glove.

This paper says: "Not so fast."

The researchers looked at forests in Southern Italy (a Mediterranean hotspot) over the last 20 years. They found that while the forests look greener and busier from space (the assembly line is running), the trees themselves are actually stopping their growth in many places. The "shipping department" is slowing down even though the "assembly line" is still chugging along.

The Analogy: The Stressed Athlete

Think of a forest tree like a marathon runner.

• The Runner's Heart Rate (Photosynthesis): Even when the runner is tired, their heart might still be beating fast because they are trying to keep going.
• The Runner's Speed (Growth): But, because they are so tired and dehydrated, they can't actually run any faster. In fact, they might be slowing down.

In the Mediterranean, the "runner" (the tree) is facing a harsh climate: hotter summers and less rain.

1. The "Greening" Illusion: Because there is more CO2 in the air (like a super-charged oxygen tank), the trees can still make energy. From a satellite, the forest looks lush and green. It looks healthy.
2. The Hidden Crisis: But the heat and drought are so intense that the trees have to stop "spending" that energy on building new wood (growth). Instead, they hoard all their energy just to survive—keeping their leaves alive and fixing damage to their water pipes.

The satellite sees the green leaves and thinks, "Great! The forest is thriving!" But the tree rings (the tree's diary) say, "I'm barely hanging on; I haven't grown a bit this year."

What the Researchers Did

They didn't just look at satellite photos (which only show the "green mask"). They used a super-smart computer model called 3D-CMCC-FEM.

Think of this model as a digital twin of the forest. It simulates the biology of the trees, the soil, and the weather. It tracks every drop of water and every gram of carbon.

• They ran the model for the last 20 years (the "Present Day").
• They also ran a "Baseline" version (what would have happened if the climate hadn't gotten so crazy).
• They compared the two.

The Key Findings

1. The Decoupling: In many parts of the forest, the link between "making energy" and "growing wood" has broken. The trees are making energy but not using it to grow.
2. Who is Suffering?
   • Deciduous Oaks (Turkey Oaks): These are the big victims. They are losing their ability to grow, even though they look green.
   • Beech Trees: At high elevations, they are surprisingly resilient (they have big energy reserves). But at lower elevations, they are struggling.
   • Evergreens (Holm Oaks & Pines): These are the tough guys. They are used to the dry heat and are handling the stress better than the others.
3. The "Legacy" Effect: The current state of the forest matters. If a forest is too crowded (too many trees fighting for water) or if the trees are young and small, they are more likely to crash under the heat.

Why Should We Care?

This is a warning sign.

If we only look at satellite images, we might think, "Oh, the forests are fine, they are getting greener!" We might stop worrying about them.
But this paper warns us: The forest is wearing a green mask.

The trees are silently starving. They are using up their savings (carbon reserves) just to stay alive. If the droughts keep getting worse, these trees might suddenly die off because they have no energy left to repair themselves.

The Takeaway

We need to stop just looking at how "green" a forest looks from space. We need to look under the hood to see if the trees are actually growing or just barely surviving. If we want to save these forests, we need to understand that looking healthy doesn't always mean being healthy. We need to manage the forests (like thinning out crowded trees) to help them survive the coming heatwaves.

Technical Summary

1. Problem Statement

Mediterranean forests are increasingly vulnerable to climate change, characterized by intensifying droughts and heatwaves. A critical challenge in monitoring forest health is the potential decoupling between canopy-level photosynthetic activity (Gross Primary Productivity, GPP) and actual tree growth (biomass accumulation).

• The Gap: Remote sensing (RS) data often show "greening" trends (increased vegetation indices), which can mask underlying physiological stress, structural decline, or mortality. Conversely, dendrochronological data (tree rings) are spatially sparse and expensive to collect.
• The Question: Does a decline in summer photosynthesis necessarily translate to a decline in tree growth, or can these processes diverge due to carbon allocation strategies and climate variability? Understanding this relationship is vital for predicting forest resilience and tipping points.

2. Methodology

The study employed a mechanistic, process-based forest ecosystem model combined with remote sensing data and statistical analysis to disentangle carbon assimilation from growth over a 20-year period (2005–2023).

• Study Area: The Basilicata region in Southern Italy (~2,400 km²), representing a heterogeneous Mediterranean landscape with diverse species (deciduous oaks, beech, pines, chestnuts) and elevations (sea level to 2,200m).
• Model Used: 3D-CMCC-FEM (Three-Dimensional Coupled Model Carbon Cycle – Forest Ecosystem Module).
  • Mechanism: The model simulates carbon and water cycles at the stand level, explicitly tracking carbon allocation from GPP to leaves, roots, and stems. It includes a carbon reserve pool that buffers stress periods, allowing trees to maintain metabolic functions even when current photosynthesis is low, often at the expense of growth.
  • Inputs: Forced by ERA5-Land meteorological data (temperature, precipitation, humidity, radiation) and atmospheric CO₂ concentrations.
  • Scenarios:
    1. Present-day: 2005–2023 using actual climate data.
    2. Baseline: A synthetic scenario created by randomly sampling years from 1980–2004 to isolate the impact of recent climate variability.
• Data Integration:
  • Remote Sensing: Landsat-derived indices (EVI, NDVI, TCG) and GOSIF-GPP (Solar-Induced Fluorescence) were used to validate model outputs and assess canopy greening trends.
  • Metrics: The study compared trends in Summer GPP (June–August) and Basal Area Increment (BAI) as a proxy for average tree growth.
• Statistical Analysis:
  • Trend analysis using the Theil–Sen estimator and Mann-Kendall test.
  • Random Forest modeling to identify the relative importance of biotic (stand density, age, DBH) and abiotic (climate, soil) drivers in explaining spatial patterns of BAI trends.

3. Key Contributions

• Decoupling Evidence: The study provides robust evidence of a decadal-scale decoupling between GPP and tree growth in Mediterranean forests. It demonstrates that canopy greening (positive GPP trends) can coexist with declining radial growth (negative BAI trends).
• Mechanistic Insight: It elucidates the role of carbon reserves in mediating this decoupling. Under recurrent drought, trees may maintain photosynthesis (or recover quickly) but divert carbon reserves to survival and hydraulic repair rather than stem growth, leading to "pre-visual" decline.
• Species and Structure Specificity: The research highlights how forest structure (density, age) and species traits (ring porosity, evergreen vs. deciduous) modulate the climate response. For instance, dense, young stands are more susceptible to decoupling than mature, high-elevation forests.
• Methodological Framework: It successfully integrates process-based modeling with RS data to overcome the spatial limitations of tree-ring data and the physiological limitations of satellite indices.

4. Key Results

• Divergent Trends:
  • Overall: While many areas showed positive GPP trends (greening), a significant portion of the forest area exhibited negative BAI trends.
  • Decoupling Classes: Under the present-day scenario, the area characterized by positive GPP but negative BAI (GPPpos-BAIneg) increased by ~3.9% compared to the baseline. Conversely, the area with coupled positive trends (GPPpos-BAIpos) decreased by ~13.1%.
• Species-Specific Responses:
  • Deciduous Oaks (Quercus cerris): Showed the most pronounced decoupling. ~25% of oak forests fell into the GPPpos-BAIneg category, indicating that despite maintaining photosynthetic activity, their growth is severely constrained.
  • Beech (Fagus sylvatica): High-elevation beech forests showed a unique pattern: negative summer GPP trends but positive BAI trends. This suggests that beech trees at higher elevations may utilize carbon reserves or benefit from earlier spring growth to sustain wood formation despite summer heat stress. Low-elevation beech, however, showed growth decline.
  • Evergreens (Holm Oak, Aleppo Pine): Generally maintained coupled positive trends, reflecting higher inherent drought tolerance.
• Drivers of Variability:
  • Random Forest Analysis: Initial stand density was the most important predictor of BAI trends, followed by age class, GPP, and spring/summer drought indices (SPEI).
  • Climate Impact: Recurrent droughts (2007, 2017, 2020, 2022) amplified the decoupling, particularly in low-to-mid elevation areas.
• Remote Sensing vs. Reality: RS data (EVI, GOSIF) largely confirmed the positive GPP trends but failed to detect the concurrent growth decline, highlighting the "masking" effect of canopy greening.

5. Significance and Implications

• Early Warning System: The study argues that relying solely on remote sensing "greening" signals is insufficient for assessing forest health. A divergence between GPP and growth serves as an early warning of structural vulnerability and potential mortality risk before canopy dieback becomes visible.
• Carbon Sink Reliability: The decoupling suggests that Mediterranean forests may be losing their capacity to sequester carbon in woody biomass even if photosynthesis remains stable, threatening the stability of the forest carbon sink.
• Management and Restoration: The findings emphasize the need for management strategies that consider forest structure (e.g., thinning dense stands to reduce competition) and species composition to enhance resilience.
• Future Research: The paper underscores the necessity of mechanistic modeling to interpret RS data and calls for improved monitoring of carbon reserves and hydraulic status to predict tipping points in forest resilience under accelerating climate change.

In conclusion, the paper reveals that Mediterranean forests are undergoing a silent transformation where photosynthetic activity no longer guarantees growth, driven by climate stress and mediated by carbon allocation strategies. This decoupling poses a significant risk to forest sustainability and carbon storage capabilities.