Aims The continuous observation datasets of water, heat, and carbon fluxes measured by the eddy covariance technique are important basis for accurate assessment of regional carbon sequestration and water-holding capacity. However, the rate of gaps in flux datasets is high and common due to various reasons, and different gap-filling methods increase the uncertainties of the related studies. The aim of this study is to introduce and test the applicability of boosted regression trees model (BRT), one of the up-to-date machine learning algorithms, for the gap- filling to flux datasets.

Methods Based on the published valid dataset of water, heat and CO₂ flux, and main environmental factors, including air temperature, atmospheric water vapor pressure, wind speed, solar shortwave radiation, topsoil temperature, and topsoil water content of an alpine Potentilla fruticosa scrubland on the northeastern Qingzang Plateau from 2003 to 2005, the BRT were trained to fill flux data gaps and the results were compared to those corresponding data serials provided by Chinese Flux Observation and Research Network (ChinaFLUX).

Important findings The results showed that the BRT performed well for a large amount of samples (N > 10 000) and the regression slopes of observation data against predicted value were between 1.01 and 1.05 with R² > 0.80. The BRT revealed that the daytime 30-min CO₂ flux (net ecosystem CO₂ exchange, NEE) in the growing season (i.e., May to October) was mainly controlled by solar shortwave radiation and atmospheric vapor pressure, whose relative contributions to NEE variability were up to 74.7%. The topsoil temperature was the determinant for NEE at night during the growing season and the whole day during the non-growing season, and its relative contribution was 68.5%. The 30-min sensible heat flux (H) and latent heat flux (LE) were both linearly related to solar radiation, and their relative contributions were above 58.6%. 30-min flux data gap amount filled by the BRT was significantly less than those by ChinaFLUX. Except for daily net ecosystem CO₂ exchange (p = 0.14), daily gross ecosystem CO₂ exchange (GEE), ecosystem respiration (RES), H, and LE of the BRT were significantly less than those of ChinaFLUX by 17.5%, 21.0%, 2.7%, and 2.2%, respectively. However, there was a reasonable consistency between the daily fluxes of 2003-2005 interpolated by the BRT and by ChinaFLUX due to the small magnitude difference (the regression slopes of the two data series were between 0.95 and 1.17). Except for monthly GEE and RES, monthly NEE, H, and LE of the BRT had no significant difference between the BRT and ChinaFLUX (p > 0.09). Compared with the ChinaFLUX gap-filling method, BRT can simulate the nonlinear relationships between fluxes and environmental factors without complicated mathematical expressions and quantify the relative contribution of environmental factors to the flux data gaps, which is a feasible technique for the integrated analysis of flux data.

Aims Eddy covariance (EC) systems are widely used for measuring the fluxes of carbon, water, and energy, as well as meteorological factors. As one important reference of independently evaluating scalar flux by EC technique, energy balance closure is widely used for evaluating data quality of carbon, water, and energy fluxes.

Methods Using the data of energy fluxes and meteorological variables retrieved from 56 site-year, the energy balance closure of six sites across three ecosystems (i.e. desert steppe, typical steppe, and meadow steppe) was analyzed by two widely used methods: linear regression from the ordinary least squares (OLS) and the energy balance ratio (EBR). The overall evaluation of energy balance closure, the seasonal and interannual variations and the related influencing factors were investigated.

Important findings The results show that: 1) the multiple-year EBR and OLS slope over the six sites had a mean value of 0.89 ± 0.11 and 0.96 ± 0.04, respectively, which are better than the results of the FLUXNET and ChinaFLUX. 2) There were significant differences over different sites and grassland types, with EBR of desert steppe (1.01 ± 0.09) and typical steppe (0.90 ± 0.11) both higher than meadow steppe (0.83 ± 0.05). There were seasonal variations of EBR over the six studied sites, and with better and stable results in growing season than non-growing season. The air temperature (T_a), vapor pressure deficit (VPD), soil moisture (SWC), and Albedo regulated the seasonal variation of EBR, with the low T_a and high Albedo remarkably reducing EBR during the non-growing season. 3) There were significant interannual variations of EBR across different sites and grassland types. The latent heat fraction (the ratio of latent heat flux to net radiation, LE/R_n), mean annual air temperature (MAT) and growing season Albedo significantly influenced interannual variation of EBR. The LE/R_n showed the strongest impact and explained 44% of the interannual variation of EBR. The significantly increasing in leaf area index (LAI) strongly regulated the upward of the available energy (net radiation minus ground heat flux, R_n- G₀), which contributes to the significant downward of EBR during observed years. It should be noted that EBR and OLS slope should be combined to better evaluate the energy balance closure. In conclusion, this study help improve our understanding of the potential linkage between energy balance closure and environmental factors, evaluate the quality of scalar flux estimates from EC technique, as well as improve the data processing protocol of flux data in the semi-arid and arid grassland region.

Aims Ecosystem apparent quantum yield (α) and maximum photosynthetic rate (P_max) are important parameters reflecting the photosynthetic characteristics of ecosystem, and also important physiological parameters in ecosystem model simulation and remote sensing inversion. The objectives of this study were to: (1) analyze the characteristics and spatial-temporal variations of the light response parameters of ecosystems in arid and semi-arid areas; and (2) reveal the key factors affecting the photosynthetic parameters and its underlying mechanisms in arid and semi-arid areas, so as to provide a scientific basis for the study of ecosystem photosynthesis and response to climate change on a regional scale.

Methods The observational fluxes and synchronous meteorological data of 9 stations in arid and semi-arid area were integrated from ChinaFLUX. The non-rectangular hyperbolic equations were used to fit the light response parameters and which influencing factors were identified by linear regression, multiple stepwise regression and path analysis.

Important findings There were obvious spatial and temporal variations in ecosystem photosynthetic parameters in arid and semi-arid areas. The photosynthetic parameters increased gradually from desert, desert grassland, typical grassland to meadow grassland. Precipitation was the dominant environmental factor of the spatial variations of photosynthetic parameters, and it also affected the spatial variation of leaf area index, both of them jointly determine the spatial variation of photosynthetic parameters. α and P_max have an obvious increasing trend with the increase of precipitation, and there was a significant negative correlation between temperature and α, but the effect of radiation on the spatial variation of photosynthetic parameters was not significant. In the growing season, P_max and α increased first and then decreased, but the monthly variability and peak time of different vegetation types were different, the photosynthetic parameters of meadow grassland had the greatest monthly variability. The monthly dynamics of α was mainly controlled by temperature and radiation, while P_max was regulated by temperature and radiation in desert and desert grassland, and by soil water content in typical grassland and meadow grassland. Ecosystem α were 0.000 47-0.002 12 mg·μmol^-1, and P_max were 0.11-0.78 mg·m^-2·s^-1 in arid and semi-arid area, which were at a low level compared with other grassland ecosystems. High temperature and low soil water supply were likely the main factors restricting the photosynthetic parameters in arid and semi-arid areas.

Aims We aimed to explore the response of net ecosystem productivity (NEP) and carbon use efficiency (CUE) to asymmetric daytime vs. nighttime warming in Artemisia ordosica shrublands, and to examine the sensitivity of carbon balance components to daytime vs. nighttime warming.

Methods The BIOME-BGC model was parameterized and validated against eddy covariance measurements of ecosystem carbon fluxes, and used for simulating the impacts of different warming scenarios on NEP and CUE and their components, including gross primary productivity (GPP), net primary productivity (NPP), ecosystem respiration (Re), autotrophic respiration (AR), heterotrophic respiration (HR), maintenance respiration (MR), and growth respiration (GR). Two warming scenarios were simulated: (1) asymmetric warming according to the historical trends from 1954 to 2020 (i.e. daytime warming 1.2 °C, nighttime warming 1.8 °C); (2) daytime or nighttime warming separately with different temperature increase treatments (2, 4, 6 °C).

Important findings (1) Modeled GPP on the daily and annual scales, Re on the daily timescale and NEP on the annual scale showed good agreement with the observed values (coefficient of determination (R²): 0.72-0.88; Nash-Sutcliffe efficiency coefficient (NS): 0.72-0.79). Modeled Re on the annual timescale and NEP on the daily timescale showed weak agreement with observed values (R²: 0.57 and 0.26; NS 0.46 and 0.12, respectively). (2) All warming scenarios promoted GPP, NPP, Re and all respiration components. GPP, Re, AR, and MR were more sensitive to daytime than to nighttime warming, while NPP, HR, GR were more sensitive to nighttime than daytime warming. (3) Greater increases in Re (about 13%) and AR (about 16%) than that in GPP (about 10%) under all warming scenarios, leading to the decreases in NEP and CUE. In addition, both NEP and CUE were more sensitive to daytime than nighttime warming. (4) NEP and CUE decreased by about 68% and 5% under the historical trend of asymmetric daytime vs. nighttime warming treatment. Greater response of NEP and CUE to the daytime warming than nighttime warming. Our results highlight the negative impacts of climatic warming on carbon sink of the semiarid shrublands, and justify the efforts to mitigate climate change are vital for dryland ecosystems.

Aims Water use efficiency (WUE) is a crucial parameter reflecting the coupling of carbon and water cycles in terrestrial ecosystems. Qingzang Plateau (QP) is the ecological barrier of China and its accommodated ecosystem is extremely sensitive to global change. Revealing the ecosystem WUE pattern and the driving forces is critical for improving our understanding on the process and mechanism of carbon and water cycles in the alpine ecosystem of the QP, which are the basis for vegetation conservation and restoration.

Methods Using the Global Land Surface Satellite (GLASS) data, meteorological data and vegetation type data, the spatio-temporal changes of WUE and their responses to temperature, precipitation, solar radiation, vapor pressure deficit (VPD), CO₂ concentration and leaf area index (LAI) during 1982-2018 over the QP were analyzed in this study. The trend magnitude and the influencing factors on WUE were further compared among vegetation types.

Important findings (1) The WUE decreases gradually from southeast to northwest on the QP, with an overall annual mean value of 1.64 g C·kg^-1. Evident differences in WUE are observed among vegetation types, with the highest value in forest and the lowest value in alpine desert. In addition, the WUE in alpine meadow is higher than that in alpine steppe. (2) The QP is prevailed by an increasing trend in WUE. Significantly increasing trends are observed in all vegetation types except for forest and cultural vegetation. Meanwhile, the variation of WUE is dominated by ecosystem gross primary productivity over 77.84% of the study area. (3) The WUE variation is mainly regulated by LAI and CO₂ concentration on the QP, and these two factors both cause positive effects on WUE. Increasing VPD inhibits WUE in alpine steppe, alpine vegetation, cultural vegetation and alpine desert.

Aims Arid and semi-arid regions are typical ecologically fragile areas, and they also have an important impact on global warming. Those regions are considered to be important CH₄ sinks since most soils are under aerobic conditions. Studies have found that along with the increase of CH₄ uptake velocity, the rate of CO₂ emissions also has increased. This study was carried out to examine whether there is an offset phenomenon and under what environmental conditions it occurs.

Methods Based on the integration of soil greenhouse gas fluxes and relevant environmental data in arid and semi-arid regions of China, correlations between soil CO₂ and soil CH₄ fluxes, on seasonal and daily scales, were analyzed.

Important findings The results showed that there were three levels of soil CO₂ and soil CH₄ flux, i.e., synergy (positively correlated), offset (negatively correlated), and random (not correlated). Among which, the proportion of random relationships was the highest, on seasonal and daily scales 83% and 54%, respectively. Compared to water content and vegetation conditions, air temperature correlated with the correlations between the two fluxes more strongly, showing a quadratic relationship (the absolute values of correlation coefficients between fluxes decreased with increasing temperature). On a seasonal scale, the mean air temperature during the sampling period determined the correlations between the fluxes with an accuracy of 92%, and the air temperature threshold of flux coupling-decoupling was 12.5 °C. On the daily scale, the diurnal air temperature difference determined fluxes relationships with an accuracy of 79% and the temperature threshold of flux coupling-decoupling was 15.2 °C. In addition, when the soil was in the state of absorbing CH₄ on a daily scale, the relationship between soil CH₄fluxand soil CO₂flux was positive in more cases. This phenomenon was difficult to explain by temperature alone. We speculate that a one-way coupling relationship between soil respiration and CH₄ oxidation formed through O₂ competition, that is, soil respiration would inhibit the CH₄ oxidation by consuming O₂, resulting in an increase in soil CO₂ emissions and a decrease in CH₄ absorption. The study suggests that coupling-decoupling of soil CO₂ and CH₄ fluxes might be driven by a mechanism of temperature regulation linked with oxygen competition regulation. Climate warming may cause decoupling of the two fluxes across space and time and increase the complexity of carbon cycles, thereby increasing the uncertainty of regional carbon flux estimations.

Aims Under the background of climate warming, the contradiction between forest and water in semi-arid areas is becoming increasingly prominent. Understanding the changes in energy flux and evapotranspiration (ET) of the plantation ecosystems in this area can provide a reference for future selection of afforestation tree species.

Methods In this paper, the water and heat fluxes of Pinus tabuliformis and Pinus sylvestris var. mongolica plantations in semi-arid area of western Liaoning were observed continuously for one year (October 2019 to October 2020) using the eddy covariance method. The length of growing season (April 11 to October 10) was determined using the critical temperature combined with the variation characteristics of normalized differential vegetation index (NDVI). The seasonal dynamics of latent heat flux (LE), sensible heat flux (H), net radiation (R_n), soil heat flux (G) and ET were analyzed. The effects of air temperature (T_a), R_n, vapor pressure deficit (VPD), soil water content (SWC), and NDVI on ET were discussed by applying regression analysis and path analysis.

Important findings The R_n, G and H of P. tabuliformis and P. sylvestrisvar. mongolica plantations showed single-peak seasonal variation trends, and the seasonal dynamics of LE fluctuated more sharply. During the whole year, energy consumption was dominated by H, followed by LE, and G consumed less energy. The average values of Bowen ratio during the growing season were 1.82 and 2.23, respectively, smaller than annual average values (3.43 and 4.44). The ET during the growing season was 302.79 and 247.54 mm, accounting for 82.89% and 84.20% of the annual ET, respectively. The annual ET was 365.29 and 293.99 mm accounting for 87.81% and 72.23% of precipitation in the same period. Priestley-Taylor coefficient (α) and decoupling factor (Ω) were used to analyze the effects of SWC and canopy conductance (g_c) on ET. The annual mean values ​​of α was 0.30 and 0.24, respectively, and the Ω values was 0.12 and 0.07, respectively. During the whole year, SWC was the dominant factor affecting the ET of two plantations, followed by R_n. Under non-water-stressed conditions, R_n had a greater impact on ET.The combined effects of T_a and VPD on ET were small, which were mostly indirect effects. NDVI and g_c were important biological factors affecting the ET of the two plantations, especially during the growing season. This study shows that P. tabuliformis and P. sylvestrisvar. mongolica plantations in the semi-arid area of western Liaoning Province adopt a conservative water consumption strategy to maintain water balance of the ecosystem, and these species are suitable afforestation tree species in this area.

Soil respiration is mainly composed of the CO₂ released from atmosphere-soil interface and change of CO₂ stored in the soil. Understanding the production and migration of CO₂ in the soil is essential for measuring the carbon cycle in terrestrial ecosystems. The flux gradient method calculates soil CO₂ flux by measuring the diffusion-driven CO₂ concentration gradient and diffusion coefficient. The flux of soil CO₂ and its stable carbon isotopes composition (δ¹³C) at different depths can be calculated based on Fickʼs law. The amount of CO₂ released from soil and the amount of CO₂ stored in different soil layers can thus be measured. The underground soil CO₂ (¹³CO₂ and ¹²CO₂) concentration is mainly controlled by pore tortuosity, the depth of root distribution, microbial activity and total soil CO₂ production. The underground CO₂ transmission process is mainly controlled by the CO₂ concentrations, porosity and water content at different depths of the soil. These physical, chemical and biological features of the soil are key factors affecting the application of the soil flux gradient method, and directly determine the precision and accuracy of soil CO₂ and its δ¹³C flux calculation. The gradient method is a useful complement to the chamber method, which can clarify the process of production and migration of soil CO₂ at different depths and thus the impacts on the release and storage of soil CO₂, elucidating the contribution of soils at different depths to CO₂ release and uncovering the underlying environmental and physical mechanisms.

Aims In recent years, the frequency of extreme drought and extreme precipitation events in arid and semi-arid regions is increasing, which has a profound impact on ecosystem productivity and carbon cycling. In desert ecosystems, vegetation is sensitive to precipitation changes, but the differences in sensitivity of aboveground productivity of ephemeral plants to extreme drought and extreme precipitation, and the moderating effect of slope positions on them need to be further studied.

Methods In this study, we experimentally reduced and increased precipitation amounts by 65% during two consecutive growing seasons in situ across four different slope positions (the bottom of the sand dune facing west (BW), the middle of the sand dune facing west (MW), the middle of the sand dune facing east (ME), and the bottom of the sand dune facing east (BE)) along sand dunes in the southern edge of the Gurbantünggüt Desert. We analyzed the sensitivity of aboveground productivity of ephemeral plants to extreme drought and extreme precipitation, and discussed the synergistic effect of slope position factors and the driving mechanism of aboveground productivity of ephemeral plants.

Important findings The results showed that: (1) Overall, the relationship between aboveground net primary productivity (ANPP) of ephemeral plants and precipitation in growing season was asymmetric, and the ANPP was more sensitive to extreme drought than to extreme precipitation. (2) Specifically, the relationship between precipitation and ANPP of ephemeral plants at BW, MW, and BE were non-linear, i.e., the rate of increase in ANPP decreases with the increased precipitation; The existence of two dominant species in the middle of the sand dune facing east changed the relationship between aboveground productivity and precipitation in the growing season, and the relationship between them showed a positive linear relationship. (3) In addition, the density of ephemeral plants (the sum of all species density) had the greatest direct impact on aboveground productivity of ephemeral plants, suggesting that extreme drought resulted in physiological death of ephemeral plants and reduced aboveground productivity by reducing plant density, while extreme precipitation increased population density to overcome meristem constraints and improve aboveground productivity. This study provides a scientific basis for accurately assessing the dynamics of carbon cycle in desert ecosystems under the background of frequent extreme climate events.

Aims Since the dynamic distribution of aquatic plants can reflect the variation of water ecological environment, it is of great significance to fully understand the spatial and temporal distribution characteristics of aquatic plants for better lake management and monitoring.

Methods On the basis of Landsat image data, this study calculated three vegetation indices, including normalized difference water index (NDWI), green veg index (Green), and macroalgae index (MAI), and constructed an aquatic plant extraction for the Donghu Lake of Wuhan using the decision tree classification method. With this method, we mapped the seasonal distribution of emergent/floating and submerged plants in the Donghu Lake in 2020, as well as their inter-annual variations during a 31-year period from 1990 to 2020.

Important findings Our results showed that the decision tree model is capable of determining the distribution of aquatic plants in the Donghu Lake accurately, with an overall accuracy of 82.29% and Kappa coefficient of 72.39%. The analysis of the seasonal variation of aquatic plants in the Donghu Lake reveal that the area of aquatic plants first increased and then decreased. In spite of being limited in February, the distribution area gradually expanded from April to August, and subsequently declined after October. The distribution and area of aquatic plants varies greatly, which can be divided into three stages regarding the long-term analysis. In the first stage (1990-1996), the area of emergent/floating plants decreased first and then increased, while that of submerged plants presented an increasing trend on a continued basis. In the second stage (1997-2015), the area of submerged plants and emergent/floating plants exhibited significant fluctuations from year to year. During this period, the area of aquatic plants reached a maximum of 2.61 km², whereas the minimum of 0.49 km². In the third stage (2016-2020), the aquatic plants in the Donghu Lake gradually recovered, leading to a 30% increase in the area of emergent/floating plants and a 18% increase in the area of submerged plants. Upon a study of the relationship between the area of aquatic plants, annual average temperature, and annual precipitation over the recent three decades, a conclusion can be drawn that annual average temperature and annual precipitation had little influence on the area of aquatic plants in the Donghu Lake. Instead, we found that environmental indicators, such as total phosphorus content, total nitrogen content, water depth, transparency, and turbidity, have significant spatial differences in the Donghu Lake, which are likely to affect the distribution of aquatic plants.

Aims Rosa roxburghii is a special economic tree species native to southwest China, and it has been taken as a key fruit tree in karst mountainous areas of Guizhou Province. However, it is unclear whether R. roxburghii can adapt to high-calcium habitats in the karst areas. To provide scientific evidence for determining the potential planting sites of R. roxburghii, the calcium-tolerance type and adaptability to high-calcium habitats of R. roxburghii were needed to be clarified.

Methods In this study, different organs from 50 R. roxburghii individuals and different calcareous soils where R. roxburghii grew were sampled in karst areas of Guizhou. The pH, exchangeable calcium and magnesium contents in the soil, total calcium and calcium oxalate contents in the plant organs, and calcium and magnesium contents in the leaves were determined, and the proportion of calcium oxalate in different organs to total calcium content was calculated. Then, the correlations between the exchangeable calcium and magnesium content in the soil with total calcium in different organs, and calcium and magnesium content in the leaves were analyzed. In addition, the distribution of calcium oxalate crystals in different organs as well as on the leaf surface was observed in samples collected from low-calcium and high-calcium habitats.

Important findings The soil at R. roxburghii sites in karst areas had a high pH, and the content of exchangeable calcium and magnesium was abundant and varies greatly. In different calcareous habitats, the total calcium content in different organs of R. roxburghii was significantly and positively correlated with the exchangeable calcium content in the soil. Similarly, the contents of total calcium and magnesium in the leaves were correlated with those in the soil. The branches and leaves were the organs that accumulated the most calcium, whereas the calcium enrichment in the roots and fruit were relatively low. The total calcium and magnesium contents of all the sample leaves ranged from 1.71% to 2.73%, among which, the proportions of calcium oxalate in branches and leaves were 55.81% and 52.76% of the total calcium, respectively; while only 29.34% and 34.30% were in roots and fruit. In samples from a high-calcium habitats, calcium oxalate prismatic crystals were abundant in the branches, leaves, and around stomatal pores, while few needle crystals were observed in the fruits. However, in samples from a low-calcium environment, no calcium oxalate crystals were observed in the roots, fruit, and leaf stomatal pores. In this study, we clarified that the R. roxburghii is an intermediate type of calcium-tolerance, which has strong adaptability to different calcium habitats in karst areas. The physiological characteristics of storing abundant calcium oxalate crystals in branches and leaves, and excreting calcium through leaf stomata play important roles in reducing the calcium concentration in R. roxburghii tissues and alleviating damages from high-calcium stress. The results of this study provide important guidance for the determination of suitable planting sites for R. roxburghii in karst areas.

Aims In recent decades, the rapid climate warming had affected chilling and heat accumulation during winter and spring and made profound changes in plant spring phenology. To date, most related studies focused on either a range of species grown in various gardens or on experimental research, which may be not necessarily applicable to real-world conditions.

Methods By using leaf-unfolding data of 25 woody species during 2003-2019 in warm temperate forests of the Dongling Mountain, Beijing, we simulated the daily chilling and heat accumulation by applying partial least square regression, dynamic model and growing degree hour model. We then analyzed the response of leaf-unfolding dates to the variation of chilling and heat accumulation by linear regressions. Finally, the differences of leaf-unfolding dates and their responses to the variation of chilling and heat accumulation between shrubs and trees were compared by ANOVA.

Important findings The chilling periods of 25 woody species were from October 6 to March 17 of next year, with the forcing periods from January 21 to April 26. The corresponding chilling and heat accumulation were 66.16 chill portion (CP) and 2 933.12 growing degree hour (GDH) on average. The leaf-unfolding dates were delayed 3.54 d per 10 CP and 7.09 d per 1 000 GDH as the chilling and heat accumulation changed, with 2 and 23 species significant, respectively. This indicated that leaf-unfolding dates of woody species in warm temperate zone were mainly affected by heat accumulation. Moreover, the leaf-unfolding dates of shrubs were 3.87 d earlier and required 543.56 GDH less heat than trees. Species leafed earlier required less heat accumulation than those leafed out later, presumably due to the opportunistic strategy adopted by shrubs and early-leafing species. Sensitivity of leaf-unfolding dates of shrubs to heat accumulation (delayed 8.10 d per 1 000 GDH) existed marginally significant difference with trees (delayed 6.13 d per 1 000 GDH), which implied that leaf-unfolding dates of shrubs might advance faster than trees as global warming progresses.