An observation-based study correlated the intensity of land cover and land use change (LCLUC) recorded in successive NLCD snapshots with land surface temperature, vapor pressure, and precipitation across conterminous U.S. ecoregions, finding regionally coherent sensitivity patterns, but it could not separate LCLUC forcing from large-scale circulation modes (ENSO, PDO, North American Monsoon) because the land-change record was decadal and the analysis correlational. The open problem is to causally attribute the observed decadal climate signals to observed LCLUC, using coupled land-atmosphere (and, where needed, ocean-coupled) counterfactual ensemble experiments driven by the now-available annual land-change record (Annual NLCD, 1985-2023), quantifying the LCLUC contribution and its uncertainty relative to internal and circulation-driven variability, and confirming, refining, or refuting the empirical sensitivity patterns as causal effects.

Background

Land cover and land use change (LCLUC) is the conversion and modification of the land surface by agriculture, forestry, urbanization, and water management. It alters climate through biogeophysical pathways: surface albedo (reflectivity), evapotranspiration and the partitioning of absorbed energy into sensible versus latent heat, and surface roughness. These pathways act locally to regionally and are distinct from the carbon-cycle (biogeochemical) effects of land change; for the conterminous United States (CONUS), Barnes and Roy (2008) quantified the radiative forcing (the perturbation of Earth's surface energy balance, expressed in watts per square meter) of contemporary LCLUC-driven albedo change, and Alkama and Cescatti (2016) mapped the surface-temperature change associated with recent global forest-cover change observationally, though by spatial covariance rather than causal attribution.

The U.S. Geological Survey's National Land Cover Database (NLCD) maps land cover for CONUS at 30-m resolution; the legacy products provided land-cover and land-change layers at multi-year to decadal epochs (e.g., 2001, 2006, 2011, 2016, 2019, 2021). Xian, Loveland, Munson, Vogelmann, Zeng, and Homer (2020) combined successive NLCD change products with gridded climate data to compute, for each ecoregion (a subdivision of CONUS with distinctive climate, vegetation, and land-use character), standardized sensitivities of land surface temperature (LST, the radiometric temperature of the ground surface as observed from satellites), vapor pressure, and precipitation to the intensity of LCLUC within 1-km grid cells. They found LST and vapor pressure to be the most sensitive variables, with persistent warming associated with LCLUC in the Appalachian and parts of the central U.S., High Plains, and northwest ecoregions and cooling in many southeast and northeast ecoregions and parts of the Great Lakes and Intermountain West, while precipitation was less sensitive.

The authors were explicit that these sensitivities are correlational. The confounders are large-scale circulation modes: ENSO (El Niño-Southern Oscillation, the tropical Pacific ocean-atmosphere mode dominating interannual climate variability over North America), the PDO (Pacific Decadal Oscillation, a North Pacific sea-surface-temperature mode with a multi-decadal imprint), and the North American Monsoon (the summer circulation regime delivering rainfall to the southwestern U.S. and Mexico). Because the NLCD change products provided only decadal snapshots, the timing of land change could not be aligned with interannual climate variability, and sparse weather-station coverage inside LCLUC-affected grids limited direct validation. The study therefore called for land-atmosphere and ocean-atmosphere coupled modeling experiments driven by satellite-derived, annual-resolution land-change inputs to move from empirical sensitivity mapping to causal attribution, and for annual LCLUC datasets to resolve the timing and sequencing of land change.

Relevant modeling precedents exist but stop short of this target. Chen and Dirmeyer (2017) showed, in global climate model simulations with prescribed historical land-use/land-cover change (LULCC, used interchangeably with LCLUC), that LULCC affects afternoon precipitation over North America. Chacón, Sushama, and Beltrami (2016) simulated the biophysical impacts of land use change over North America with the Canadian Regional Climate Model — a regional climate model (RCM), i.e., a limited-area atmospheric model driven at its lateral boundaries by a reanalysis (a physically consistent gridded reconstruction of past weather, produced by assimilating historical observations into a fixed numerical model) or by global-model output. Alter, Douglas, Winter, and Eltahir (2018) attributed twentieth-century summer climate change in the central United States to agricultural intensification, including irrigation, using observational and modeling evidence — a region- and process-specific attribution rather than a CONUS-wide accounting of observed decadal LCLUC. The CORDEX (Coordinated Regional climate Downscaling Experiment, a program of the World Climate Research Programme coordinating regional climate modeling worldwide) Flagship Pilot Study LUCAS (Land Use and Climate Across Scales) established a coordinated multi-model RCM protocol contrasting idealized forestation and grassland land-cover states, with first ensemble results over Europe (Davin et al. 2020); published LUCAS results address Europe under idealized scenarios, not the observed U.S. land-change record. On the input side, the gap that motivated the source study has since closed: USGS released Annual NLCD Collection 1.0 in 2024, providing annual land cover and land change for CONUS from 1985 through 2023 (U.S. Geological Survey 2025), which removes the decadal-snapshot limitation and makes the attribution problem posed here tractable.

Problem Statement

Determine, by causal attribution rather than correlation, how much of the observed decadal-scale variation of near-surface climate — land surface temperature, vapor pressure, and precipitation — across the conterminous United States is attributable to observed land cover and land use change, with the LCLUC-induced signal quantitatively separated from internal variability (the unforced, chaotic fluctuation of the coupled ocean-atmosphere system) and from the large-scale circulation modes (ENSO, the Pacific Decadal Oscillation, and the North American Monsoon) that confounded the empirical analysis of Xian et al. (2020).

A complete answer should: (1) drive coupled land-atmosphere (and, where the question requires, ocean-coupled) ensemble experiments over CONUS with the observed annual-resolution land-change chronology — now available as Annual NLCD 1985-2023 — as the land boundary forcing, against a counterfactual ensemble identical in every respect except that land cover is held fixed at a baseline epoch; (2) quantify, at the ecoregion scale, the LCLUC-attributable contribution to each climate variable and its uncertainty, using a defensible separation of the land-forcing signal from internal variability and from circulation-mode variability; and (3) state explicitly which of the empirical sensitivity patterns reported by Xian et al. (2020) — the regional warming and cooling patterns, the vapor-pressure increases in the Great Plains and West Coast, the weak precipitation sensitivity — are confirmed, refined, or contradicted as causal effects of LCLUC, together with the physical mechanisms responsible. A rigorous demonstration that the LCLUC-attributable signal is statistically indistinguishable from zero at the achievable detection power (the smallest true signal the experimental design could reliably detect) would equally resolve the question by refuting the causal interpretation of those empirical sensitivities.

References

  1. George Z. Xian, Thomas Loveland, Seth M. Munson, James E. Vogelmann, Xubin Zeng, Collin J. Homer (2020). Climate sensitivity to decadal land cover and land use change across the conterminous United States. Global and Planetary Change 192, 103262. DOI: 10.1016/j.gloplacha.2020.103262. https://doi.org/10.1016/j.gloplacha.2020.103262
  2. U.S. Geological Survey (2025). Annual NLCD (National Land Cover Database)—The next generation of land cover mapping. U.S. Geological Survey Fact Sheet 2025-3001. DOI: 10.3133/fs20253001. https://doi.org/10.3133/fs20253001
  3. Edouard L. Davin, Diana Rechid, Marcus Breil, Rita M. Cardoso, Erika Coppola, Peter Hoffmann, Lisa L. Jach, Eleni Katragkou, Nathalie de Noblet-Ducoudré, Kai Radtke, Mario Raffa, Pedro M. M. Soares, Giannis Sofiadis, Susanna Strada, Gustav Strandberg, Merja H. Tölle, Kirsten Warrach-Sagi, Volker Wulfmeyer (2020). Biogeophysical impacts of forestation in Europe: first results from the LUCAS (Land Use and Climate Across Scales) regional climate model intercomparison. Earth System Dynamics 11(1), 183-200. DOI: 10.5194/esd-11-183-2020. https://doi.org/10.5194/esd-11-183-2020
  4. Liang Chen, Paul A. Dirmeyer (2017). Impacts of Land-Use/Land-Cover Change on Afternoon Precipitation over North America. Journal of Climate 30(6), 2121-2140. DOI: 10.1175/JCLI-D-16-0589.1. https://doi.org/10.1175/jcli-d-16-0589.1
  5. Arlette Chacón, Laxmi Sushama, Hugo Beltrami (2016). Biophysical Impacts of Land Use Change over North America as Simulated by the Canadian Regional Climate Model. Atmosphere 7(3), 34. DOI: 10.3390/atmos7030034. https://doi.org/10.3390/atmos7030034
  6. Ross E. Alter, Hunter C. Douglas, Jonathan M. Winter, Elfatih A. B. Eltahir (2018). Twentieth Century Regional Climate Change During the Summer in the Central United States Attributed to Agricultural Intensification. Geophysical Research Letters 45(3), 1586-1594. DOI: 10.1002/2017GL075604. https://doi.org/10.1002/2017GL075604
  7. Christopher A. Barnes, David P. Roy (2008). Radiative forcing over the conterminous United States due to contemporary land cover land use albedo change. Geophysical Research Letters 35(9). DOI: 10.1029/2008GL033567. https://doi.org/10.1029/2008GL033567
  8. Ramdane Alkama, Alessandro Cescatti (2016). Biophysical climate impacts of recent changes in global forest cover. Science 351(6273), 600-604. DOI: 10.1126/science.aac8083. https://doi.org/10.1126/science.aac8083
  9. Pedro Vertino de Queiroz, Rezaul Mahmood, Lilyan Fulginiti, Richard Perrin (2025). Land use policies, land use land cover change, and climate: a review with examples from the United States. Environment, Development and Sustainability. DOI: 10.1007/s10668-025-06023-8. https://doi.org/10.1007/s10668-025-06023-8
  10. Dávid D. Kovács, Eatidal Amin, Katja Berger, Pablo Reyes-Muñoz, Jochem Verrelst (2023). Untangling the Causal Links between Satellite Vegetation Products and Environmental Drivers on a Global Scale by the Granger Causality Method. Remote Sensing 15(20), 4956. DOI: 10.3390/rs15204956. https://doi.org/10.3390/rs15204956