<?xml version='1.0' encoding='utf-8'?>
<?xml-stylesheet type="text/xsl" href="/v2/static/oai2.xsl"?>
<OAI-PMH xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd">
  <responseDate>2026-10-14T19:05:18Z</responseDate>
  <request identifier="oai:figshare.com:article/32051055" metadataPrefix="oai_dc" verb="GetRecord">https://api.figshare.com/v2/oai</request>
  <GetRecord>
    <record>
      <header>
        <identifier>oai:figshare.com:article/32051055</identifier>
        <datestamp>2026-10-08T12:53:24Z</datestamp>
        <setSpec>category_25939</setSpec>
        <setSpec>item_type_3</setSpec>
        <setSpec>month_year_10_2026</setSpec>
      </header>
      <metadata>
        <oai_dc:dc xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"  xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns:dc="http://purl.org/dc/elements/1.1/" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd">
          <dc:title>Declining but spatially redistributed global fire nitrogen emissions regulated by vegetation-climate interactive mechanisms</dc:title>
          <dc:creator>Yuan Jiang (23752413)</dc:creator>
          <dc:creator>Wei Song (30780)</dc:creator>
          <dc:creator>Xue-Yan Liu (2021578)</dc:creator>
          <dc:subject>Earth system sciences</dc:subject>
          <dc:subject>Fire emission</dc:subject>
          <dc:description>&lt;p dir="ltr"&gt;This R script implements a four-parameter multiplicative uncertainty framework for estimating gridded inorganic nitrogen (N) emissions from open biomass burning, combining emission factors (EFs), combustion factors (CFs), satellite-derived burned area (BA), and aboveground biomass (AGB/fuel load). Uncertainty is quantified through two complementary approaches: first-order error propagation for analytical per-pixel uncertainty and parameter contribution breakdown, and Monte Carlo simulation with Latin Hypercube Sampling (LHS) featuring region-specific EF sampling, block-LHS spatial correlation structure, and freeze-one contribution decomposition. The framework covers seven biomes (boreal, temperate, and tropical forest; grassland; shrubland; savanna; cropland), seven N species (NH3, NO, NO2, HNO2, HNO3, NH4+, NO3-), and continental aggregation across six continents.&lt;/p&gt;&lt;p dir="ltr"&gt;A complementary machine learning (ML) module predicts emissions from multi-source environmental drivers using three models: LightGBM, XGBoost, and Elastic Net, evaluated via dual cross-validation (random stratified and spatial block) with grid search hyperparameter tuning. Emissions are detrended pixel-wise (Theil-Sen) prior to training and the trend surface is restored to preserve physical non-negativity and long-term trends. Interpretability is provided through SHAP values, partial dependence plots, and standardized coefficients, with performance assessed via R², RMSE, MAE, NSE, KGE, and bias on a 2017–2020 held-out test set. Moran’s I diagnosis quantifies spatial autocorrelation to justify spatial block CV.&lt;/p&gt;</dc:description>
          <dc:date>2026-04-27T15:12:52Z</dc:date>
          <dc:type>Dataset</dc:type>
          <dc:type>Dataset</dc:type>
          <dc:identifier>10.6084/m9.figshare.32051055.v7</dc:identifier>
          <dc:relation>https://figshare.com/articles/dataset/Productivity_drives_global_biomass_burning_nitrogen_emissions/32051055</dc:relation>
          <dc:rights>CC BY 4.0</dc:rights>
        </oai_dc:dc>
      </metadata>
    </record>
  </GetRecord>
</OAI-PMH>
