Vegetation index (VI) time series derived from satellite sensors have been widely used in the estimation of vegetation parameters, but the quality of VI time series is easily affected by clouds and poor atmospheric conditions. The function fitting method is a widely used effective noise reduction technique for VI time series, but it is vulnerable to noise. Thus, ancillary data about VI quality are utilized to alleviate the interference of noise. However, this approach is limited by the availability, accuracy, and application rules of ancillary data. In this paper, we aimed to develop a new reconstruction method that does not require ancillary data. Based on the assumptions that VI time series follow the gradual growth and decline pattern of vegetation dynamics, and that clouds or poor atmospheric conditions usually depress VI values, we proposed a reconstruction method for VI time series based on self-weighting function fitting from curve features (SWCF). SWCF consists of two major procedures: (1) determining a fitting weight for each VI point based on the curve features of the VI time series and (2) implementing the weighted function fitting to reconstruct the VI time series. The double logistic function, double Gaussian function, and polynomial function were tested in SWCF based on a simulated dataset. The results indicate that the weighted function fitting with SWCF outperformed the corresponding unweighted function fitting with the root-mean-square error (RMSE) significantly reduced by 26.82–52.44% (p < 0.05), and it also outperformed the Savitzky–Golay filtering with the RMSE significantly reduced by 13.98–54.04% (p < 0.05) for 270 sample points selected in mid-high latitudes of the Northern Hemisphere. Moreover, SWCF showed excellent robustness and applicability in regional applications.

The light use efficiency (LUE) model has been widely used in regional and global terrestrial gross primary productivity (GPP) estimation due to its simple structure, few input parameters, and particular theoretical basis. As a key input parameter of the LUE model, the maximum LUE (Ɛ_max) is crucial for the accurate estimation of GPP and to the interpretability of the LUE model. Currently, most studies have assumed Ɛ_max as a universal constant or constants depending on vegetation type, which means that the spatiotemporal dynamics of Ɛ_max were ignored, leading to obvious uncertainties in LUE-based GPP estimation. Using quality-screened daily data from the FLUXNET 2015 dataset, this paper proposed a photosynthetically active radiation (PAR)-regulated dynamic Ɛ_max (PAR-Ɛ_max, corresponding model named PAR-LUE) by considering the nonlinear response of vegetation photosynthesis to solar radiation. The PAR-LUE was compared with static Ɛ_max-based (MODIS and EC-LUE) and spatial dynamics Ɛ_max-based (D-VPM) models at 171 flux sites. Validation results showed that (1) R² and RMSE between PAR-LUE GPP and observed GPP were 0.65 (0.44) and 2.55 (1.82) g C m⁻² MJ⁻¹ d⁻¹ at the 8-day (annual) scale, respectively; (2) GPP estimation accuracy of PAR-LUE was higher than that of other LUE-based models (MODIS, EC-LUE, and D-VPM), specifically, R² increased by 29.41%, 2.33%, and 12.82%, and RMSE decreased by 0.36, 0.14, and 0.34 g C m⁻² MJ⁻¹ d⁻¹ at the annual scale; and (3) specifically, compared to the static Ɛ_max-based model (MODIS and EC-LUE), PAR-LUE effectively relieved the underestimation of high GPP. Overall, the newly developed PAR-Ɛ_max provided an estimation method utilizing a spatiotemporal dynamic Ɛ_max, which effectively reduced the uncertainty of GPP estimation and provided a new option for the optimization of Ɛ_max in the LUE model.

Pollen allergies have a serious impact on people’s physical and mental health. Accurate and efficient prediction of the outbreak date of pollen allergies plays an important role in the conservation of people sensitive to allergenic pollen. It is a frontier research to combine new social media data and satellite data to develop a model to forecast the outbreak date of pollen allergies. This study extracted the real outbreak dates of spring pollen allergies from Sina Weibo records from 2011 to 2021 in Beijing and calculated five vegetation indices of three vegetation types as phenological characteristics within the 30 days before the average outbreak date. The sensitivity coefficients and correlation coefficients were used to screen the phenological characteristics that best reflected the outbreak date of spring pollen allergy. Based on the best characteristic, two kinds of prediction models for the outbreak date of spring pollen allergy in Beijing were established (the linear fit prediction model and the cumulative linear fit prediction model), and the root mean square error (RMSE) was calculated as the prediction accuracy. The results showed that (1) the date of EVI2 (2-band enhanced vegetation index) in evergreen forest first reaching 0.138 can best reflect the outbreak date of pollen allergies in spring, and (2) the cumulative linear fit prediction model based on EVI2 in evergreen forests can obtain a high accuracy with an average RMSE of 3.6 days, which can predict the outbreak date of spring pollen allergies 30 days in advance. Compared with the existing indirect prediction models (which predict pollen concentrations rather than pollen allergies), this model provides a new direct way to predict pollen allergy outbreaks by using only remote sensing time-series data before pollen allergy outbreaks. The new prediction model also has better representativeness and operability and is capable of assisting public health management.

Accurate estimation of terrestrial ecosystem respiration (TER) is essential for refining global carbon budgets. Current large-scale TER models rely on empirical structures derived from site-scale observations, often driven solely by hydrothermal factors. However, incorporating ecosystem-scale information is critical for more accurate large-scale TER modeling. Such ecosystem-scale variables have not been well parameterized, since the mechanisms by which they affect TER remain unclear. To address this gap, here we developed a Causality constrained Interpretable Machine Learning model for TER estimation (named “CIML-TER”) which consider the ecosystem-scale information. CIML-TER exhibited higher estimation accuracy (reducing relative mean absolute error by approximately 15%) and overcame the “artificial discontinuities” phenomenon of traditional models. Meanwhile, we quantitatively revealed that although environmental factors, such as temperature and water, were still the dominant drivers of TER (contributing ~44.15% of global TER variability), biotic factors (e.g., vegetation structure, ~25.91%) and spatiotemporal variation factors (e.g., land cover and phenology, ~29.94%) were also critical.