Modeling Nutrition Quality and Storage of Forage Using Climate Data and Normalized-Difference Vegetation Index in Alpine Grasslands
1
Lhasa Plateau Ecosystem Research Station, Key Laboratory of Ecosystem Network Observation and Modeling, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing 100101, China
2
Key Laboratory of Digital Earth Science, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China
3
State Key Laboratory of Remote Sensing Science, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China
*
Author to whom correspondence should be addressed.
Remote Sens. 2022, 14(14), 3410; https://doi.org/10.3390/rs14143410
Submission received: 23 June 2022
/
Revised: 8 July 2022
/
Accepted: 14 July 2022
/
Published: 15 July 2022
Abstract
:Quantifying forage nutritional quality and pool at various spatial and temporal scales are major challenges in quantifying global nitrogen and phosphorus cycles, and the carrying capacity of grasslands. In this study, we modeled forage nutrition quality and storage using climate data under fencing conditions, and using climate data and a growing-season maximum normalized-difference vegetation index under grazing conditions based on four different methods (i.e., multiple linear regression, random-forest models, support-vector machines and recursive-regression trees) in the alpine grasslands of Tibet. Our results implied that random-forest models can have greater potential ability in modeling forage nutrition quality and storage than the other three methods. The relative biases between simulated nutritional quality using random-forest models and the observed nutritional quality, and between simulated nutrition storage using random-forest models and the observed nutrition storage, were lower than 2.00% and 6.00%, respectively. The RMSE between simulated nutrition quality using random-forest models and the observed nutrition quality, and between simulated nutrition storage using random-forest models and the observed nutrition storage, were no more than 0.99% and 4.50 g m−2, respectively. Therefore, random-forest models based on climate data and/or the normalized-difference vegetation index can be used to model forage nutrition quality and storage in the alpine grasslands of Tibet.
1. Introduction
Forage nutritional quality and pool can affect the quality and size of livestock and wildlife, and the nutrient-carrying capacity in various grassland ecosystems [1,2,3]. Crude protein (CP), ether extract (EE), crude ash (Ash), acid detergent fiber (ADF), neutral detergent fiber (NDF) and water-soluble carbohydrate (WSC) contents and pools are often treated as indicators of forage nutritional quality and nutritional pools, respectively [4,5]. Quantifying their variations at various spatial and temporal scales are major challenges in quantifying the global nitrogen and phosphorus cycles and carrying capacity of grasslands [2,6]. More and more studies have estimated forage nutritional quality and pool at various spatial and temporal scales [2,7,8,9,10]. However, there are still some uncertainties. First, compared to plant biomass and/or production, forage nutritional quality and pool have been less modeled at various spatial and temporal scales [11,12]. Second, compared to CP, the other five variables (i.e., EE, Ash, ADF, NDF and WSC) have been less quantified by previous studies [13,14]. Meanwhile, CP by itself cannot fully reflect changes in forage nutrition quality and pool [1,4,15,16]. Third, many of these previous studies have modeled forage nutrition quality and/or pool using only remote-sensing data [9,13,14], because plant growth conditions can be indicated by remote-sensing data and can be closely related to forage nutritional quality and pools [4,17]. However, remote-sensing data are concurrently influenced by climate change and humanoid activities [18,19]. For example, climate change can directly affect various vegetation indices (e.g., the normalized-difference vegetation index) [20,21,22,23]. Grazing can also directly affect the normalized-difference vegetation index and the soil-adjusted vegetation index [24]. Moreover, it is tricky to accurately separate the effects of climate change and humanoid activities on terrestrial ecosystems, especially at bulky scales [11]. These findings indicate that it is hard to examine the exclusivity relationships between forage nutrition and climate variables based on simulated forage nutrition from remote-sensing data alone. Fourth, previous studies have mainly estimated the nutrient quality and pool of forage via exponential regression, stepwise linear regression, partial least-squares regression and the derivative transformation technique for hyperspectral data [25,26,27]. As an increasingly mature technology, big-data-mining technology (e.g., random-forest models, support-vector machines and recursive-regression trees) has higher precision and data processing ability than other methods [28]. However, to the best of our knowledge, only a few studies have tried to simulate the nutritional quality and pool of plants using big-data-mining technology. That is, it remains unclear whether or not big-data-mining technology can be used to simulate forage nutritional quality and pools. Moreover, it is still not clear whether the accuracy of big-data-mining technology in simulating forage nutritional quality and pools is greater than that of other methods. Therefore, more studies are needed to better predict forage nutritional quality and pool at various spatial and temporal scales in grasslands around the world under global change.
Forage nutritional quality and pool can determine the quality and size of livestock and wildlife and affect the policy of balance between forage and livestock in alpine grasslands on the Tibetan Plateau. At present, the policy of balance between forage and livestock mainly depends on forage yield rather than nutritional pool on the Tibetan Plateau. However, amount carrying capacity always overestimates or underestimates the nutritional carrying capacity in alpine grasslands on the Tibetan Plateau [29,30]. These findings imply that a combination of amount carrying capacity and nutritional carrying capacity can be better in the accuracy quantization of carrying capacity. The accurate quantification of forage nutritional quality and pool is an important basis for the accurate quantification of livestock carrying capacity and the balance between forage and livestock in alpine grasslands on the Tibetan Plateau. It is expected that big-data-mining technology (e.g., random-forest models) may be used to simulate forage nutritional quality and pool, whereas, to our best knowledge, few studies have tried to evaluate and compare the accuracies of different methods of big-data-mining technology in predicting forage CP, EE, Ash, ADF, NDF and WSC contents and pools in alpine grasslands on the Tibetan Plateau. Therefore, it is still not clear which one is the better big-data-mining technology in predicting the nutrient quality and pools of forage in alpine grasslands on the Tibetan Plateau.
Random-forest models, support-vector machines and recursive-regression trees are three commonly used big-data-mining technologies [28], and they have been developed as packages (i.e., randomForest, e1071 and rpart) of the R software. The packages randomForest, e1071 and rpart can greatly facilitate our processing of big data and the simulation/prediction of forage nutritional quality and pool. Therefore, in this study, we modeled forage nutritional quality and pool in alpine grasslands from climate data and/or the growing-season maximum normalized-difference vegetation index under fencing or grazing conditions in Tibet, based on multiple linear regressions, random-forest models, support-vector machines and recursive-regression trees. The main objectivities of this study were to compare the accuracies of the random-forest models, support-vector machines and recursive-regression trees in predicting forage CP, EE, Ash, ADF, NDF and WSC contents and pools in alpine grasslands.
2. Materials and Methods
2.1. Plant Sampling and Analyses
In July–August 2018–2020, we clipped the aboveground biomass of all plants from 190 quadrats under fencing conditions and 190 quadrats under grazing conditions in alpine grasslands of Tibet. The sampling sites are illustrated in Figure 1. The quadrat sizes were 0.50 m × 0.50 m and 1.00 m × 1.00 m for alpine meadows and alpine steppes, respectively. The aboveground biomass was weighed after oven-drying at 65 °C for 48 h. Then, we measured the CP, EE, Ash, ADF/NDF and WSC using the Kjeldahl method, Soxhlet extraction method, complete combustion method, Van Soest method and anthrone-based method, respectively [4].
2.2. Normalized-Difference Vegetation Index and Climate Data
The growing-season (May–September) maximum normalized-difference vegetation index was obtained using a Moderate-Resolution Imaging Spectroradiometer (MOD13A3, Collection 6, 1 km × 1 km, monthly). Monthly air temperature, precipitation and radiation data were obtained from interpolated climate data with a spatial resolution of 1 km × 1 km [31]. According to previous studies [11,32], we assumed that the CP, EE, Ash, ADF, NDF and WSC contents and pools under fencing conditions were only affected by climate change (air temperature, precipitation and radiation), and they had potential forage nutritional quality and pools. By contrast, we assumed the CP, EE, Ash, ADF, NDF and WSC contents and pools under grazing conditions were simultaneously affected by climate change and human activities, and they had actual nutritional quality and pool. Growing-season mean air temperature, total precipitation and total radiation were used to simulate the potential CP, EE, Ash, ADF, NDF and WSC contents and pools under fencing conditions based on multiple linear regressions, random-forest models, support-vector machines and recursive-regression trees, respectively (Table 1, Table 2, Table 3 and Table 4). By contrast, growing-season mean air temperature, total precipitation, total radiation and the maximum normalized-difference vegetation index were used to simulate the actual CP, EE, Ash, ADF, NDF and WSC contents and pools under grazing conditions based on multiple linear regressions, random-forest models, support-vector machines and recursive-regression trees, respectively (Table 1, Table 2, Table 3 and Table 4).
2.3. Statistical Analysis
We used the cross-validations across all 190 samples under fencing or grazing conditions. The 190 samples were randomly divided into two groups. The first group (n = 170) was used to obtain the multiple linear regressions, random-forest models, support-vector machines and recursive-regression trees for each of the six variables under fencing or grazing conditions. The second group (n = 20) was used to validate the multiple linear regressions, random-forest models, support-vector machines and recursive-regression trees. The relative bias, root-mean-square error (RMSE), relative RMSE, determination coefficient (R2) and linear slope between simulated and observed data were treated as indicators of model accuracies. The multiple linear regressions, random-forest models, support-vector machines and recursive-regression trees were performed using R.4.1.2.
3. Results
3.1. Model Building
The key parameters of the models between the nutritional quality and pool variables (i.e., CP, EE, Ash, ADF, NDF and WSC contents and pools) and growing-season climate data (i.e., mean air temperature, total precipitation and total radiation), and/or the maximum normalized-difference vegetation index under fencing or grazing conditions, are shown in Table 1, Table 2, Table 3 and Table 4, respectively. Different methods can provide different parameters among the four methods (Table 1, Table 2, Table 3 and Table 4). For example, the multiple linear regressions, random-forest models, and recursive-regression trees can directly provide R2 values (Table 1, Table 2 and Table 4). By contrast, the support-vector machines did not directly provide R2 values (Table 3). Among the multiple linear regressions, random-forest models, and recursive-regression trees, the multiple linear regressions explained the fewest variations in all the nutritional quality and pools variables, and the random-forest models explained the most variations for most of the nutritional quality and pools variables (Table 1, Table 2 and Table 4). The explanation abilities of the environmental variables of forage nutritional quality and pool can change with different indices of forage nutritional quality and pool, and land use types (Table 1, Table 2 and Table 4). Climate data can explain about 73–93%, 5–75% and 24–92% of the variation in these variables related to forage nutritional quality and pool, based on the random-forest models, multiple linear regressions and recursive-regression trees under fencing conditions, respectively (Table 1, Table 2 and Table 4). Meanwhile, climate data and growing-season maximum normalized-difference vegetation index can, together, explain about 76–96%, 8–57% and 42–91% of the variation in these variables related to forage nutritional quality and pool, based on the random-forest models, multiple linear regressions and recursive-regression trees under grazing conditions, respectively (Table 1, Table 2 and Table 4). No fixed or default ntree and mtry parameters were used for the random-forest models (Table 2). Additionally, no fixed support-vector parameter was used for the support-vector machines (Table 3).
3.2. Model Validation
The RMSE and relative RMSE values between the simulated and observed nutritional quality and pool variables under fencing or grazing conditions are shown in Table 5 and Table 6. The RMSE and relative RMSE values between the simulated nutritional quality and pool variables using random-forest models, and the observed nutritional quality and pool variables, were the lowest among the four simulated methods under both fencing and grazing conditions, respectively. The RMSE and relative RMSE values between simulated nutrition quality using random-forest models and observed nutrition quality were no more than 0.99% and 7.23%, respectively. Meanwhile, the RMSE and relative RMSE values between the simulated nutrition storage using random-forest models and observed nutrition storage were no more than 4.50 g m−2 and 35.32%, respectively.
The relative biases between the simulated and observed nutritional quality and pool variables under fencing or grazing conditions are shown in Table 7. The absolute values of the relative biases between the simulated—using multiple linear regressions—and observed Ash contents, EE contents, ADF contents and NDF pools were the largest among the four simulated methods under both fencing and grazing conditions. The absolute values of the relative biases between the simulated—using support-vector machines—and observed CP pools and WSC pools were the largest among the four simulated methods under both fencing and grazing conditions. The absolute values of the relative biases between the simulated—using recursive-regression trees—and observed WSC contents were the largest among the four simulated methods under both fencing and grazing conditions. The absolute values of the relative biases between the simulated—using random-forest models—and observed Ash pools, EE contents, WSC pools and NDF pools were the lowest among the four simulated methods under both fencing and grazing conditions. All the relative biases between the simulated and observed nutritional quality and pool variables were lower than 6.00% for the random-forest models, whereas those of the other three methods were greater than 6.00% (even 20.00%) for some cases.
The linear slopes between the simulated and observed nutritional quality and pool variables under fencing or grazing conditions are shown in Table 8. Generally, the slopes between the simulated forage nutritional quality and pool variables using random-forest models, and the observed forage nutritional quality and pool variables, were the closest to 1 compared to the other three methods. The linear slopes were within 0.55–1.12, 0.94–1.02, 0.47–1.04 and 0.77–1.30 between the observed and simulated nutritional quality and pool variables from the multiple linear regressions, random-forest models, support-vector machine and recursive-regression trees, respectively.
The R2 values between the simulated and observed nutritional quality and pool variables under fencing or grazing conditions are shown in Figure 2, Figure 3, Figure 4, Figure 5, Figure 6, Figure 7, Figure 8 and Figure 9. Generally, the R2 values between the simulated forage nutritional quality and pool variables using random-forest models, and the observed forage nutritional quality and pool variables, were the closest to 100% compared to the other three methods. The simulated nutritional quality and pool variables from multiple linear regressions, random-forest models, support-vector machine and recursive-regression trees can explain 47–100%, 93–100%, 56–100% and 57–100% of the observed nutritional quality and pool variables, respectively.
4. Discussion
The R2 values of the random-forest models, multiple linear regression and recursive-regression trees were within 0.73–0.96, 0.05–0.75 and 0.24–0.92, respectively (Table 1, Table 2 and Table 4), indicating that generally, random-forest models can have a greater explanatory ability in forage nutritional quality and pools than multiple linear regression and recursive-regression trees. Our findings indicated that the R2 values of random-forest models under grazing conditions were generally greater than those under fencing conditions (Table 2). That is, the combination of climate data and growing-season maximum normalized-difference vegetation index may better model CP, EE, Ash, ADF, NDF and WSC contents and pools in the alpine grasslands of Tibet under grazing conditions. However, the data sources of the random-forest models under fencing conditions are different from those under grazing conditions in this study. Moreover, the R2 values of the multiple linear regression and recursive-regression trees under grazing conditions were lower than those under fencing conditions for some cases (Table 1 and Table 4). Therefore, further studies are needed on whether or not the combination of climate data and the growing-season maximum normalized-difference vegetation index had closer relationships with forage nutritional quality and pool than single climate data in the alpine grasslands of Tibet.
Climate data had closer relationships with forage nutritional quality than nutritional storage under fencing conditions. Moreover, the predicted accuracies of forage nutritional pools were lower than those of forage nutritional quality in most cases. These findings may be related to the fact that forage nutritional pools were equal to the multiplication of forage nutritional quality and aboveground plant biomass. As is well known, aboveground plant biomass is generally directly related to air temperature, precipitation and radiation in alpine grasslands on the Tibetan Plateau [12,20,21,23,33,34]. However, to the best of our knowledge, climate data alone cannot capture 100% of the variation in aboveground plant biomass in alpine grasslands on the Tibetan Plateau [12,31]. On the other hand, both the explained ability of climate data in aboveground plant biomass and the predicted accuracy of aboveground plant biomass models derived from climate data may be lower than those of nutritional quality in alpine grasslands on the Tibetan Plateau [28].
Our findings implied that different methods can have different predicted accuracies, which may be mainly due to their different algorithms. For example, firstly, the correlations between the dependent variables and independent variables are assumed to be linear for the multiple-linear-regression method. In contrast, the random-forest models cannot directly assume the linear or nonlinear relationships among independent variables, nor their linear or nonlinear relationships with dependent variables. Actually, the forage nutritional quality variables cannot always have linear relationships with climate data [4,12,35]; this, in turn, may result in the lower predicted accuracies of forage nutritional quality and pool for the multiple linear regressions, but the higher predicted accuracies of forage nutritional quality and pool for the random-forest models in this study. Secondly, ntree and mtry are two important and key parameters of random-forest models, and they are generally not fixed values. Users can try their best to find the relatively optimal combination of ntree and mtry by adjusting the parameters of ntree and mtry, which can guarantee that a relatively optimal random-forest model is obtained. Similarly, the number of support vectors, as one key parameter of a support-vector machine, is not generally a fixed value. By contrast, there are no adjustable parameters for the multiple-linear-regression method. Thirdly, randomness is one obvious characteristic of the random-forest model, but not for the multiple-linear-regression method.
Our findings suggest that random-forest models can have the greatest potential ability to predict forage nutritional quality and pools in alpine grasslands among the four methods. The predicted accuracies of the CP, EE, Ash, ADF, NDF and WSC contents and pools from the random-forest models in this study were greater than those reported by previous studies [2,25,36,37,38,39]. For example, a previous simulated study demonstrated that the relative biases of CP, EE, ADF and NDF were 5.97%, 2.30%, 3.82% and 3.82% in alpine grasslands of the Qilian Mountain, respectively [36], which were greater than those in this study (Table 5). The R2 values were 0.23–0.76 for the CP pool of model building in alpine grasslands of the Haibei region [36], which were lower than the R2 values (0.83–0.93) in this study. Moreover, in this study, the default parameters of ntree and mtry were not used, whereas the relatively optimal combination of ntree and mtry parameters was used. Therefore, the random-forest models established by this study can be used to predict changes in forage nutritional quality and pool under climate change and grazing conditions in alpine grasslands on the Tibetan Plateau.
5. Conclusions
In this study, we established and evaluated four methods (i.e., random-forest models, multiple linear regression, support-vector machines and recursive-regression trees) for the CP, EE, Ash, ADF, NDF and WSC contents and pools in the alpine grasslands of Tibet, under fencing or grazing conditions, respectively. The predicted accuracies of these random-forest models were relatively higher than those of the other three methods. The simulated nutritional quality using random-forest models can have high accuracies, with relative biases of <2.00% and an RMSE of <0.99%. The simulated nutritional storage using random-forest models can also have high accuracies, with relative biases of <6.00% and an RMSE of <4.50 g m−2. The linear slopes were within 0.55–1.12, 0.94–1.02, 0.47–1.04 and 0.77–1.30 between the observed and simulated nutritional quality and pool variables using multiple linear regressions, random-forest models, the support-vector machine and recursive-regression trees, respectively. The simulated nutritional quality and pool variables from multiple linear regressions, random-forest models, the support-vector machine and recursive-regression trees can explain 47–100%, 93–100%, 56–100% and 57–100% of the observed nutritional quality and pool variables, respectively. Therefore, the established random-forest models can be used to model nutritional quality and storage in alpine grasslands on the Tibetan Plateau.
Author Contributions
Conceptualization, G.F. and F.H.; methodology, G.F.; software, C.Y.; validation, G.F.; formal analysis, G.F. and F.H.; investigation, G.F.; resources, C.Y.; data curation, G.F.; writing—original draft preparation, G.F. and F.H.; writing—review and editing, G.F. and F.H.; visualization, S.W.; supervision, S.W.; project administration, C.Y.; funding acquisition, C.Y. All authors have read and agreed to the published version of the manuscript.
Funding
This research was funded by the Youth Innovation Promotion Association of the Chinese Academy of Sciences [2020054]; the Second Comprehensive Scientific Expedition on the Tibetan Plateau [2019QZKK0302]; the National Natural Science Foundation of China [31600432]; the Bingwei Outstanding Young Talents Program of the Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences [2018RC202]; the Tibet Science and Technology Major Projects of the Pratacultural Industry [XZ202101ZD0003N]; the Science and Technology Project of the Tibet Autonomous Region [XZ202101ZD0007G, XZ202201ZY0003N]; the STS Project of the Chinese Academy of Sciences [KFJ-STS-QYZD-2021-22-003]; the construction of the fixed Observation and Experimental Station of the first Support System for Agricultural Green Development in Zhongba County; and the Central Government Guides Local Science and Technology Development Program [XZ202202YD0009C].
Data Availability Statement
The datasets presented in this article are not readily available because the datasets generated for this study are only available on request to the corresponding author.
Acknowledgments
We appreciate the editor and reviewers for their profound and professional comments which greatly improved this study.
Conflicts of Interest
The authors declare no conflict of interest.
References
- Fu, G.; Wang, J.; Li, S.; He, P. Responses of forage nutrient quality to grazing in the alpine grassland of Northern Tibet. Acta Prataculturae Sin. 2021, 30, 38–50. [Google Scholar]
- Wang, X.; Liu, S.-J.; Jia, H.-F.; Chai, S.-T.; Dang, A.-R.; Liu, X.-H.; Hao, L.-Z.; Cui, Z.-H. Study on the Nutrition of Alpine Meadow Based on Hyperspectral Data. Spectrosc. Spectr. Anal. 2012, 32, 2780–2784. [Google Scholar] [CrossRef]
- Craine, J.M.; Elmore, A.; Angerer, J.P. Long-term declines in dietary nutritional quality for North American cattle. Environ. Res. Lett. 2017, 12, 044019. [Google Scholar] [CrossRef] [Green Version]
- Sun, W.; Li, S.; Zhang, Y.; Fu, G. Effect of long-term experimental warming on the nutritional quality of alpine meadows in the Northern Tibet. J. Resour. Ecol. 2020, 11, 516–524. [Google Scholar]
- Fanselow, N.; Schonbach, P.; Gong, X.Y.; Lin, S.; Taube, F.; Loges, R.; Pan, Q.M.; Dittert, K. Short-term regrowth responses of four steppe grassland species to grazing intensity, water and nitrogen in Inner Mongolia. Plant Soil 2011, 340, 279–289. [Google Scholar] [CrossRef]
- Dale, L.M.; Pierna, J.A.F.; Vermeulen, P.; Lecler, B.; Bogdan, A.D.; Pacurar, F.S.; Rotar, I.; Thewis, A.; Baeten, V. Research on crude protein and digestibility of Arnica montana L. using conventional NIR spectrometry and hyperspectral imaging NIR. J. Food Agric. Environ. 2012, 10, 391–396. [Google Scholar]
- Zengeya, F.M.; Mutanga, O.; Murwira, A. Linking remotely sensed forage quality estimates from WorldView-2 multispectral data with cattle distribution in a savanna landscape. Int. J. Appl. Earth Obs. Geoinf. 2013, 21, 513–524. [Google Scholar] [CrossRef]
- Hogrefe, K.R.; Patil, V.P.; Ruthrauff, D.R.; Meixell, B.W.; Budde, M.E.; Hupp, J.W.; Ward, D.H. Normalized difference vegetation index as an estimator for abundance and quality of Avian herbivore forage in Arctic Alaska. Remote Sens. 2017, 9, 1234. [Google Scholar] [CrossRef] [Green Version]
- Thoma, D.P.; Bailey, D.W.; Long, D.S.; Nielsen, G.A.; Henry, M.P.; Breneman, M.C.; Montagne, C. Short-term monitoring of rangeland forage conditions with AVHRR imagery. J. Range Manag. 2002, 55, 383–389. [Google Scholar] [CrossRef]
- Beeri, O.; Phillips, R.; Hendrickson, J.; Frank, A.B.; Kronberg, S. Estimating forage quantity and quality using aerial hyperspectral imagery for northern mixed-grass prairie. Remote Sens. Environ. 2007, 110, 216–225. [Google Scholar] [CrossRef]
- Sun, W.; Li, M.; Wang, J.H.; Fu, G. Driving mechanism of gross primary production changes and implications for grassland management on the Tibetan Plateau. J. Resour. Ecol. 2019, 10, 472–480. [Google Scholar]
- Wang, J.; Li, M.; Yu, C.; Fu, G. The change in environmental variables linked to climate change has a stronger effect on aboveground net primary productivity than does phenological change in alpine grasslands. Front. Plant Sci. 2022, 12, 798633. [Google Scholar] [CrossRef] [PubMed]
- Kawamura, K.; Akiyama, T.; Yokota, H.; Tsutsumi, M.; Yasuda, T.; Watanabe, O.; Wang, G.; Wang, S. Monitoring of forage conditions with MODIS imagery in the Xilingol steppe, Inner Mongolia. Int. J. Remote Sens. 2005, 26, 1423–1436. [Google Scholar] [CrossRef]
- Knox, N.M.; Skidmore, A.K.; Prins, H.H.T.; Heitkonig, I.M.A.; Slotow, R.; van der Waal, C.; de Boer, W.F. Remote sensing of forage nutrients: Combining ecological and spectral absorption feature data. Isprs J. Photogramm. Remote Sens. 2012, 72, 27–35. [Google Scholar] [CrossRef]
- Ren, H.Y.; Han, G.D.; Schonbach, P.; Gierus, M.; Taube, F. Forage nutritional characteristics and yield dynamics in a grazed semiarid steppe ecosystem of Inner Mongolia, China. Ecol. Indic. 2016, 60, 460–469. [Google Scholar] [CrossRef]
- Sasaki, T.; Ohkuro, T.; Jamsran, U.; Takeuchi, K. Changes in the herbage nutritive value and yield associated with threshold responses of vegetation to grazing in Mongolian rangelands. Grass Forage Sci. 2012, 67, 446–455. [Google Scholar] [CrossRef]
- Fu, G.; Sun, W. Temperature sensitivities of vegetation indices and aboveground biomass are primarily linked with warming magnitude in high-cold grasslands. Sci. Total Environ. 2022, 843. [Google Scholar] [CrossRef]
- Shen, Z.X.; Fu, G.; Yu, C.Q.; Sun, W.; Zhang, X.Z. Relationship between the growing season maximum enhanced vegetation index and climatic factors on the Tibetan Plateau. Remote Sens. 2014, 6, 6765–6789. [Google Scholar] [CrossRef] [Green Version]
- Wang, S.H.; Sun, W.; Li, S.W.; Shen, Z.X.; Fu, G. Interannual variation of the growing season maximum normalized difference vegetation index, MNDVI, and its relationship with climatic factors on the Tibetan Plateau. Pol. J. Ecol. 2015, 63, 424–439. [Google Scholar] [CrossRef]
- Zhang, G.; Shen, Z.; Fu, G. Function diversity of soil fungal community has little exclusive effects on the response of aboveground plant production to experimental warming in alpine grasslands. Appl. Soil Ecol. 2021, 168, 104153. [Google Scholar] [CrossRef]
- Fu, G.; Zhang, H.R.; Sun, W. Response of plant production to growing/non-growing season asymmetric warming in an alpine meadow of the Northern Tibetan Plateau. Sci. Total Environ. 2019, 650, 2666–2673. [Google Scholar] [CrossRef] [PubMed]
- Wang, J.W.; Fu, G.; Zhang, G.Y.; Shen, Z.X. The effect of higher warming on vegetation indices and biomass production is dampened by greater drying in an alpine meadow on the Northern Tibetan Plateau. J. Resour. Ecol. 2017, 8, 104–111. [Google Scholar]
- Fu, G.; Shen, Z.X.; Zhang, X.Z. Increased precipitation has stronger effects on plant production of an alpine meadow than does experimental warming in the Northern Tibetan Plateau. Agric. For. Meteorol. 2018, 249, 11–21. [Google Scholar] [CrossRef]
- Fu, G.; Shen, Z.X. Clipping has stronger effects on plant production than does warming in three alpine meadow sites on the Northern Tibetan Plateau. Sci. Rep. 2017, 7, 16330. [Google Scholar] [CrossRef] [Green Version]
- Zhang, A.; Guo, C.; Yan, W. Improving remote sensing estimation accuracy of pasture crude protein content by interval analysis. Trans. Chin. Soc. Agric. Eng. 2018, 34, 149–156. [Google Scholar]
- Gao, R.; Li, Z.D.; Ma, Z.; Kong, Q.M.; Rizwan, M.; Su, Z.B. Research on Crude Protein of Pasture Based on Hyperspectral Imaging. Spectrosc. Spectr. Anal. 2019, 39, 3245–3250. [Google Scholar] [CrossRef]
- Durante, M.; Oesterheld, M.; Pineiro, G.; Vassallo, M.M. Estimating forage quantity and quality under different stress and senescent biomass conditions via spectral reflectance. Int. J. Remote Sens. 2014, 35, 2963–2981. [Google Scholar] [CrossRef]
- Zhang, X.Z.; Li, M.; Wu, J.S.; He, Y.T.; Niu, B. Alpine Grassland Aboveground Biomass and Theoretical Livestock Carrying Capacity on the Tibetan Plateau. J. Resour. Ecol. 2022, 13, 129–141. [Google Scholar]
- Sun, P.; Cui, Z.; Liu, S.; Chai, S.; Hao, L.; Wang, X. Seasonal evaluation of nutrition and carrying capacity of grazing pastures in the Three-River Source Region. Acta Prataculturae Sin. 2015, 24, 92–101. [Google Scholar]
- Hao, L.; Han, X.; Niu, J.; Zhang, X.; Xiang, Y.; Wang, X.; Chai, S.; Long, R.; Liu, S. Study on Nutrient Balance of Grassland in Henan County, Sanjiang. Acta Agrestia Sin. 2018, 26, 520–524. [Google Scholar]
- Fu, G.; Sun, W.; Li, S.W.; Zhang, J.; Yu, C.Q.; Shen, Z.X. Modeling aboveground biomass using MODIS images and climatic data in grasslands on the Tibetan Plateau. J. Resour. Ecol. 2017, 8, 42–49. [Google Scholar]
- Chen, B.X.; Zhang, X.Z.; Tao, J.; Wu, J.S.; Wang, J.S.; Shi, P.L.; Zhang, Y.J.; Yu, C.Q. The impact of climate change and anthropogenic activities on alpine grassland over the Qinghai-Tibet Plateau. Agric. For. Meteorol. 2014, 189, 11–18. [Google Scholar] [CrossRef]
- Wang, J.; Yu, C.; Fu, G. Warming reconstructs the elevation distributions of aboveground net primary production, plant species and phylogenetic diversity in alpine grasslands. Ecol. Indic. 2021, 133, 108355. [Google Scholar] [CrossRef]
- Wu, J.S.; Fu, G. Modelling aboveground biomass using MODIS FPAR/LAI data in alpine grasslands of the Northern Tibetan Plateau. Remote Sens. Lett. 2018, 9, 150–159. [Google Scholar] [CrossRef]
- Fu, G.; Shen, Z.-X. Asymmetrical warming of growing/non-growing season increases soil respiration during growing season in an alpine meadow. Sci. Total Environ. 2022, 812, 152591. [Google Scholar] [CrossRef]
- Yao, X.; Sun, H.; Li, C.; Zhang, Z.; Chen, S.; Xie, J. Prediction Model Establishment of Near Infrared Reflectance Spectroscopy for Nutrient Quality of Natural Forage in Alpine Grassland. Chin. J. Anim. Nutr. 2021, 33, 4088–4097. [Google Scholar]
- Xue, Z.; Liu, N.; Zhang, Y. Nutritional Quality and Digestibility Evaluation of Alfalfa Hay Bale by Near Infrared Reflectance. Acta Agrestia Sin. 2017, 25, 165–171. [Google Scholar]
- Yu, L.; Wang, X.; Chai, S.; Liu, S. Seasonal Dynamic Inversion of Natural Grassland Forage Content Based on High-resolution Satellite Remote Sensing. Acta Agrestia Sin. 2020, 28, 547–557. [Google Scholar]
- Zhang, A.; Yan, W.; Guo, C. Inversion model of pasture crude protein content based on hyperspectral image. Trans. Chin. Soc. Agric. Eng. 2018, 34, 188–194. [Google Scholar]
Figure 1.
Sampling sites.
Figure 2.
Relationships (a) between simulated and observed crude protein (CP) content; (b) between simulated and observed crude ash (Ash) content; (c) between simulated and observed ether extract (EE) content; (d) between simulated and observed water-soluble carbohydrate (WSC) content; (e) between simulated and observed acid detergent fiber (ADF) content; (f) between simulated and observed neutral detergent fiber (NDF) content; (g) between simulated and observed crude protein (CP) pool; (h) between simulated and observed crude ash (Ash) pool; (i) between simulated and observed ether extract (EE) pool; (j) between simulated and observed water-soluble carbohydrate (WSC) pool; (k) between simulated and observed acid detergent fiber (ADF) pool; and (l) between simulated and observed neutral detergent fiber (NDF) pool under fencing conditions. The solid lines indicate fitted lines. All the simulated data were based on multiple linear regression.
Figure 2.
Relationships (a) between simulated and observed crude protein (CP) content; (b) between simulated and observed crude ash (Ash) content; (c) between simulated and observed ether extract (EE) content; (d) between simulated and observed water-soluble carbohydrate (WSC) content; (e) between simulated and observed acid detergent fiber (ADF) content; (f) between simulated and observed neutral detergent fiber (NDF) content; (g) between simulated and observed crude protein (CP) pool; (h) between simulated and observed crude ash (Ash) pool; (i) between simulated and observed ether extract (EE) pool; (j) between simulated and observed water-soluble carbohydrate (WSC) pool; (k) between simulated and observed acid detergent fiber (ADF) pool; and (l) between simulated and observed neutral detergent fiber (NDF) pool under fencing conditions. The solid lines indicate fitted lines. All the simulated data were based on multiple linear regression.
Figure 3.
Relationships (a) between simulated and observed crude protein (CP) content; (b) between simulated and observed crude ash (Ash) content; (c) between simulated and observed ether extract (EE) content; (d) between simulated and observed water-soluble carbohydrate (WSC) content; (e) between simulated and observed acid detergent fiber (ADF) content; (f) between simulated and observed neutral detergent fiber (NDF) content; (g) between simulated and observed crude protein (CP) pool; (h) between simulated and observed crude ash (Ash) pool; (i) between simulated and observed ether extract (EE) pool; (j) between simulated and observed water-soluble carbohydrate (WSC) pool; (k) between simulated and observed acid detergent fiber (ADF) pool; and (l) between simulated and observed neutral detergent fiber (NDF) pool under grazing conditions. The solid lines indicate fitted lines. All the simulated data were based on multiple linear regression.
Figure 3.
Relationships (a) between simulated and observed crude protein (CP) content; (b) between simulated and observed crude ash (Ash) content; (c) between simulated and observed ether extract (EE) content; (d) between simulated and observed water-soluble carbohydrate (WSC) content; (e) between simulated and observed acid detergent fiber (ADF) content; (f) between simulated and observed neutral detergent fiber (NDF) content; (g) between simulated and observed crude protein (CP) pool; (h) between simulated and observed crude ash (Ash) pool; (i) between simulated and observed ether extract (EE) pool; (j) between simulated and observed water-soluble carbohydrate (WSC) pool; (k) between simulated and observed acid detergent fiber (ADF) pool; and (l) between simulated and observed neutral detergent fiber (NDF) pool under grazing conditions. The solid lines indicate fitted lines. All the simulated data were based on multiple linear regression.
Figure 4.
Relationships (a) between simulated and observed crude protein (CP) content; (b) between simulated and observed crude ash (Ash) content; (c) between simulated and observed ether extract (EE) content; (d) between simulated and observed water-soluble carbohydrate (WSC) content; (e) between simulated and observed acid detergent fiber (ADF) content; (f) between simulated and observed neutral detergent fiber (NDF) content; (g) between simulated and observed crude protein (CP) pool; (h) between simulated and observed crude ash (Ash) pool; (i) between simulated and observed ether extract (EE) pool; (j) between simulated and observed water-soluble carbohydrate (WSC) pool; (k) between simulated and observed acid detergent fiber (ADF) pool; and (l) between simulated and observed neutral detergent fiber (NDF) pool under fencing conditions. The solid lines indicate fitted lines. All the simulated data were based on random-forest model.
Figure 4.
Relationships (a) between simulated and observed crude protein (CP) content; (b) between simulated and observed crude ash (Ash) content; (c) between simulated and observed ether extract (EE) content; (d) between simulated and observed water-soluble carbohydrate (WSC) content; (e) between simulated and observed acid detergent fiber (ADF) content; (f) between simulated and observed neutral detergent fiber (NDF) content; (g) between simulated and observed crude protein (CP) pool; (h) between simulated and observed crude ash (Ash) pool; (i) between simulated and observed ether extract (EE) pool; (j) between simulated and observed water-soluble carbohydrate (WSC) pool; (k) between simulated and observed acid detergent fiber (ADF) pool; and (l) between simulated and observed neutral detergent fiber (NDF) pool under fencing conditions. The solid lines indicate fitted lines. All the simulated data were based on random-forest model.
Figure 5.
Relationships (a) between simulated and observed crude protein (CP) content; (b) between simulated and observed crude ash (Ash) content; (c) between simulated and observed ether extract (EE) content; (d) between simulated and observed water-soluble carbohydrate (WSC) content; (e) between simulated and observed acid detergent fiber (ADF) content; (f) between simulated and observed neutral detergent fiber (NDF) content; (g) between simulated and observed crude protein (CP) pool; (h) between simulated and observed crude ash (Ash) pool; (i) between simulated and observed ether extract (EE) pool; (j) between simulated and observed water-soluble carbohydrate (WSC) pool; (k) between simulated and observed acid detergent fiber (ADF) pool; and (l) between simulated and observed neutral detergent fiber (NDF) pool under grazing conditions. The solid lines indicate fitted lines. All the simulated data were based on random-forest model.
Figure 5.
Relationships (a) between simulated and observed crude protein (CP) content; (b) between simulated and observed crude ash (Ash) content; (c) between simulated and observed ether extract (EE) content; (d) between simulated and observed water-soluble carbohydrate (WSC) content; (e) between simulated and observed acid detergent fiber (ADF) content; (f) between simulated and observed neutral detergent fiber (NDF) content; (g) between simulated and observed crude protein (CP) pool; (h) between simulated and observed crude ash (Ash) pool; (i) between simulated and observed ether extract (EE) pool; (j) between simulated and observed water-soluble carbohydrate (WSC) pool; (k) between simulated and observed acid detergent fiber (ADF) pool; and (l) between simulated and observed neutral detergent fiber (NDF) pool under grazing conditions. The solid lines indicate fitted lines. All the simulated data were based on random-forest model.
Figure 6.
Relationships (a) between simulated and observed crude protein (CP) content; (b) between simulated and observed crude ash (Ash) content; (c) between simulated and observed ether extract (EE) content; (d) between simulated and observed water-soluble carbohydrate (WSC) content; (e) between simulated and observed acid detergent fiber (ADF) content; (f) between simulated and observed neutral detergent fiber (NDF) content; (g) between simulated and observed crude protein (CP) pool; (h) between simulated and observed crude ash (Ash) pool; (i) between simulated and observed ether extract (EE) pool; (j) between simulated and observed water-soluble carbohydrate (WSC) pool; (k) between simulated and observed acid detergent fiber (ADF) pool; and (l) between simulated and observed neutral detergent fiber (NDF) pool under fencing conditions. The solid lines indicate fitted lines. All the simulated data were based on support-vector machines.
Figure 6.
Relationships (a) between simulated and observed crude protein (CP) content; (b) between simulated and observed crude ash (Ash) content; (c) between simulated and observed ether extract (EE) content; (d) between simulated and observed water-soluble carbohydrate (WSC) content; (e) between simulated and observed acid detergent fiber (ADF) content; (f) between simulated and observed neutral detergent fiber (NDF) content; (g) between simulated and observed crude protein (CP) pool; (h) between simulated and observed crude ash (Ash) pool; (i) between simulated and observed ether extract (EE) pool; (j) between simulated and observed water-soluble carbohydrate (WSC) pool; (k) between simulated and observed acid detergent fiber (ADF) pool; and (l) between simulated and observed neutral detergent fiber (NDF) pool under fencing conditions. The solid lines indicate fitted lines. All the simulated data were based on support-vector machines.
Figure 7.
Relationships (a) between simulated and observed crude protein (CP) content; (b) between simulated and observed crude ash (Ash) content; (c) between simulated and observed ether extract (EE) content; (d) between simulated and observed water-soluble carbohydrate (WSC) content; (e) between simulated and observed acid detergent fiber (ADF) content; (f) between simulated and observed neutral detergent fiber (NDF) content; (g) between simulated and observed crude protein (CP) pool; (h) between simulated and observed crude ash (Ash) pool; (i) between simulated and observed ether extract (EE) pool; (j) between simulated and observed water-soluble carbohydrate (WSC) pool; (k) between simulated and observed acid detergent fiber (ADF) pool; and (l) between simulated and observed neutral detergent fiber (NDF) pool under grazing conditions. The solid lines indicate fitted lines. All the simulated data were based on support-vector machines.
Figure 7.
Relationships (a) between simulated and observed crude protein (CP) content; (b) between simulated and observed crude ash (Ash) content; (c) between simulated and observed ether extract (EE) content; (d) between simulated and observed water-soluble carbohydrate (WSC) content; (e) between simulated and observed acid detergent fiber (ADF) content; (f) between simulated and observed neutral detergent fiber (NDF) content; (g) between simulated and observed crude protein (CP) pool; (h) between simulated and observed crude ash (Ash) pool; (i) between simulated and observed ether extract (EE) pool; (j) between simulated and observed water-soluble carbohydrate (WSC) pool; (k) between simulated and observed acid detergent fiber (ADF) pool; and (l) between simulated and observed neutral detergent fiber (NDF) pool under grazing conditions. The solid lines indicate fitted lines. All the simulated data were based on support-vector machines.
Figure 8.
Relationships (a) between simulated and observed crude protein (CP) content; (b) between simulated and observed crude ash (Ash) content; (c) between simulated and observed ether extract (EE) content; (d) between simulated and observed water-soluble carbohydrate (WSC) content; (e) between simulated and observed acid detergent fiber (ADF) content; (f) between simulated and observed neutral detergent fiber (NDF) content; (g) between simulated and observed crude protein (CP) pool; (h) between simulated and observed crude ash (Ash) pool; (i) between simulated and observed ether extract (EE) pool; (j) between simulated and observed water-soluble carbohydrate (WSC) pool; (k) between simulated and observed acid detergent fiber (ADF) pool; and (l) between simulated and observed neutral detergent fiber (NDF) pool under fencing conditions. The solid lines indicate fitted lines. All the simulated data were based on recursive-regression trees.
Figure 8.
Relationships (a) between simulated and observed crude protein (CP) content; (b) between simulated and observed crude ash (Ash) content; (c) between simulated and observed ether extract (EE) content; (d) between simulated and observed water-soluble carbohydrate (WSC) content; (e) between simulated and observed acid detergent fiber (ADF) content; (f) between simulated and observed neutral detergent fiber (NDF) content; (g) between simulated and observed crude protein (CP) pool; (h) between simulated and observed crude ash (Ash) pool; (i) between simulated and observed ether extract (EE) pool; (j) between simulated and observed water-soluble carbohydrate (WSC) pool; (k) between simulated and observed acid detergent fiber (ADF) pool; and (l) between simulated and observed neutral detergent fiber (NDF) pool under fencing conditions. The solid lines indicate fitted lines. All the simulated data were based on recursive-regression trees.
Figure 9.
Relationships (a) between simulated and observed crude protein (CP) content; (b) between simulated and observed crude ash (Ash) content; (c) between simulated and observed ether extract (EE) content; (d) between simulated and observed water-soluble carbohydrate (WSC) content; (e) between simulated and observed acid detergent fiber (ADF) content; (f) between simulated and observed neutral detergent fiber (NDF) content; (g) between simulated and observed crude protein (CP) pool; (h) between simulated and observed crude ash (Ash) pool; (i) between simulated and observed ether extract (EE) pool; (j) between simulated and observed water-soluble carbohydrate (WSC) pool; (k) between simulated and observed acid detergent fiber (ADF) pool; and (l) between simulated and observed neutral detergent fiber (NDF) pool under grazing conditions. The solid lines indicate fitted lines. All the simulated data were based on recursive-regression trees.
Figure 9.
Relationships (a) between simulated and observed crude protein (CP) content; (b) between simulated and observed crude ash (Ash) content; (c) between simulated and observed ether extract (EE) content; (d) between simulated and observed water-soluble carbohydrate (WSC) content; (e) between simulated and observed acid detergent fiber (ADF) content; (f) between simulated and observed neutral detergent fiber (NDF) content; (g) between simulated and observed crude protein (CP) pool; (h) between simulated and observed crude ash (Ash) pool; (i) between simulated and observed ether extract (EE) pool; (j) between simulated and observed water-soluble carbohydrate (WSC) pool; (k) between simulated and observed acid detergent fiber (ADF) pool; and (l) between simulated and observed neutral detergent fiber (NDF) pool under grazing conditions. The solid lines indicate fitted lines. All the simulated data were based on recursive-regression trees.
Table 1.
Regression parameters of multiple-linear-regression models for acid detergent fiber (ADF), crude ash (Ash), crude protein (CP), ether extract (EE), neutral detergent fiber (NDF) and water-soluble carbohydrate (WSC) contents and pools under fencing or grazing conditions, respectively.
Table 1.
Regression parameters of multiple-linear-regression models for acid detergent fiber (ADF), crude ash (Ash), crude protein (CP), ether extract (EE), neutral detergent fiber (NDF) and water-soluble carbohydrate (WSC) contents and pools under fencing or grazing conditions, respectively.
| Variables | Land-Use Types | Regression Coefficients | R2 | ||||
|---|---|---|---|---|---|---|---|
| Intercept | Temperature | Precipitation | Radiation | NDVI | |||
| ADF content | Fencing | −9.56 | −0.39 | 0.00 | 0.01 | 0.26 | |
| Grazing | −13.03 | 1.60 | 0.02 | 0.01 | −2.09 | 0.25 | |
| ADF pool | Fencing | −24.26 | 1.96 | 0.03 | 0.01 | 0.05 | |
| Grazing | −143.96 | 1.58 | 0.02 | 0.03 | 69.08 | 0.39 | |
| Ash content | Fencing | −55.08 | 0.80 | −0.01 | 0.02 | 0.70 | |
| Grazing | −35.98 | 1.33 | 0.01 | 0.01 | −11.00 | 0.42 | |
| Ash pool | Fencing | −49.26 | 0.67 | 0.01 | 0.01 | 0.20 | |
| Grazing | −129.79 | 1.18 | 0.00 | 0.03 | 43.03 | 0.28 | |
| CP content | Fencing | 36.37 | −0.02 | 0.00 | −0.01 | 0.75 | |
| Grazing | 42.75 | −0.59 | −0.01 | −0.01 | 1.12 | 0.57 | |
| CP pool | Fencing | 20.33 | 0.36 | 0.00 | −0.01 | 0.20 | |
| Grazing | −12.70 | 0.15 | 0.00 | 0.00 | 19.28 | 0.57 | |
| EE content | Fencing | 7.47 | −0.02 | 0.00 | 0.00 | 0.57 | |
| Grazing | 8.02 | −0.19 | 0.00 | 0.00 | −0.48 | 0.50 | |
| EE pool | Fencing | 2.95 | 0.06 | 0.00 | 0.00 | 0.06 | |
| Grazing | −2.25 | 0.00 | 0.00 | 0.00 | 3.98 | 0.44 | |
| NDF content | Fencing | 88.26 | −0.20 | 0.01 | −0.01 | 0.25 | |
| Grazing | 33.80 | 0.62 | 0.02 | 0.00 | −9.44 | 0.08 | |
| NDF pool | Fencing | 7.60 | 2.29 | 0.05 | −0.01 | 0.11 | |
| Grazing | −257.39 | 2.45 | 0.03 | 0.06 | 121.09 | 0.49 | |
| WSC content | Fencing | −2.44 | 0.18 | 0.00 | 0.00 | 0.35 | |
| Grazing | −12.70 | 0.18 | 0.00 | 0.00 | 2.82 | 0.28 | |
| WSC pool | Fencing | −4.36 | 0.19 | 0.00 | 0.00 | 0.09 | |
| Grazing | −36.28 | 0.43 | 0.00 | 0.01 | 20.99 | 0.41 | |
Table 2.
Parameters of random-forest models for acid detergent fiber (ADF), crude ash (Ash), crude protein (CP), ether extract (EE), neutral detergent fiber (NDF) and water-soluble carbohydrate (WSC) contents and pools under fencing or grazing conditions, respectively.
Table 2.
Parameters of random-forest models for acid detergent fiber (ADF), crude ash (Ash), crude protein (CP), ether extract (EE), neutral detergent fiber (NDF) and water-soluble carbohydrate (WSC) contents and pools under fencing or grazing conditions, respectively.
| Variables | Land-Use Types | R2 | Mean Square Errors | ntree | mtry |
|---|---|---|---|---|---|
| ADF content | Fencing | 0.82 | 3.82 | 773 | 3 |
| Grazing | 0.90 | 4.59 | 872 | 2 | |
| ADF pool | Fencing | 0.73 | 90.31 | 813 | 3 |
| Grazing | 0.91 | 27.89 | 897 | 4 | |
| Ash content | Fencing | 0.93 | 3.76 | 549 | 1 |
| Grazing | 0.94 | 3.09 | 887 | 4 | |
| Ash pool | Fencing | 0.78 | 13.79 | 786 | 1 |
| Grazing | 0.76 | 54.90 | 450 | 3 | |
| CP content | Fencing | 0.92 | 0.28 | 508 | 1 |
| Grazing | 0.93 | 0.40 | 788 | 1 | |
| CP pool | Fencing | 0.83 | 1.55 | 945 | 1 |
| Grazing | 0.93 | 1.29 | 340 | 3 | |
| EE content | Fencing | 0.80 | 0.04 | 501 | 2 |
| Grazing | 0.91 | 0.04 | 503 | 2 | |
| EE pool | Fencing | 0.79 | 0.08 | 331 | 2 |
| Grazing | 0.88 | 0.10 | 809 | 2 | |
| NDF content | Fencing | 0.86 | 3.19 | 824 | 3 |
| Grazing | 0.93 | 3.17 | 939 | 4 | |
| NDF pool | Fencing | 0.78 | 64.46 | 328 | 1 |
| Grazing | 0.94 | 51.90 | 755 | 4 | |
| WSC content | Fencing | 0.82 | 0.34 | 640 | 2 |
| Grazing | 0.96 | 0.23 | 850 | 4 | |
| WSC pool | Fencing | 0.77 | 0.65 | 215 | 1 |
| Grazing | 0.87 | 4.14 | 622 | 3 |
Table 3.
Parameters of support-vector machines for acid detergent fiber (ADF), crude ash (Ash), crude protein (CP), ether extract (EE), neutral detergent fiber (NDF) and water-soluble carbohydrate (WSC) contents and pools under fencing or grazing conditions, respectively.
Table 3.
Parameters of support-vector machines for acid detergent fiber (ADF), crude ash (Ash), crude protein (CP), ether extract (EE), neutral detergent fiber (NDF) and water-soluble carbohydrate (WSC) contents and pools under fencing or grazing conditions, respectively.
| Variables | Land-Use Types | Mean Residuals | Mean Decision Values | No of Support Vectors |
|---|---|---|---|---|
| ADF content | Fencing | −0.06 | 0.01 | 145 |
| Grazing | −0.32 | 0.05 | 144 | |
| ADF pool | Fencing | 5.01 | −0.27 | 173 |
| Grazing | 1.21 | −0.07 | 127 | |
| Ash content | Fencing | −0.20 | 0.03 | 138 |
| Grazing | −0.16 | 0.02 | 138 | |
| Ash pool | Fencing | 1.20 | −0.15 | 140 |
| Grazing | 2.88 | −0.19 | 162 | |
| CP content | Fencing | 0.08 | −0.04 | 126 |
| Grazing | 0.16 | −0.07 | 129 | |
| CP pool | Fencing | 0.50 | −0.17 | 138 |
| Grazing | 0.21 | −0.05 | 120 | |
| EE content | Fencing | 0.03 | −0.07 | 143 |
| Grazing | 0.01 | −0.02 | 131 | |
| EE pool | Fencing | 0.15 | −0.24 | 145 |
| Grazing | 0.13 | −0.14 | 143 | |
| NDF content | Fencing | 0.19 | −0.04 | 150 |
| Grazing | −0.22 | 0.03 | 131 | |
| NDF pool | Fencing | 3.20 | −0.19 | 145 |
| Grazing | 1.88 | −0.07 | 120 | |
| WSC content | Fencing | −0.04 | 0.03 | 155 |
| Grazing | 0.11 | −0.05 | 144 | |
| WSC pool | Fencing | 0.30 | −0.18 | 150 |
| Grazing | 0.71 | −0.13 | 137 |
Table 4.
Parameters of recursive-regression trees for acid detergent fiber (ADF), crude ash (Ash), crude protein (CP), ether extract (EE), neutral detergent fiber (NDF) and water-soluble carbohydrate (WSC) contents and pools under fencing or grazing conditions, respectively.
Table 4.
Parameters of recursive-regression trees for acid detergent fiber (ADF), crude ash (Ash), crude protein (CP), ether extract (EE), neutral detergent fiber (NDF) and water-soluble carbohydrate (WSC) contents and pools under fencing or grazing conditions, respectively.
| Variables | Land-Use Types | R2 | Minsplit | Minbucket | Maxcompete | Maxsurrogate | Usesurrogate | Surrogatestyle | Maxdepth | xval |
|---|---|---|---|---|---|---|---|---|---|---|
| ADF content | Fencing | 0.60 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 |
| Grazing | 0.84 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 | |
| ADF pool | Fencing | 0.24 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 |
| Grazing | 0.87 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 | |
| Ash content | Fencing | 0.88 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 |
| Grazing | 0.69 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 | |
| Ash pool | Fencing | 0.50 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 |
| Grazing | 0.55 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 | |
| CP content | Fencing | 0.92 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 |
| Grazing | 0.86 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 | |
| CP pool | Fencing | 0.37 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 |
| Grazing | 0.89 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 | |
| EE content | Fencing | 0.81 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 |
| Grazing | 0.84 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 | |
| EE pool | Fencing | 0.32 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 |
| Grazing | 0.77 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 | |
| NDF content | Fencing | 0.73 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 |
| Grazing | 0.42 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 | |
| NDF pool | Fencing | 0.45 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 |
| Grazing | 0.91 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 | |
| WSC content | Fencing | 0.65 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 |
| Grazing | 0.54 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 | |
| WSC pool | Fencing | 0.36 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 |
| Grazing | 0.66 | 20 | 7 | 4 | 5 | 2 | 0 | 30 | 10 |
Table 5.
The RMSE (g m−2) between simulated and observed nutritional quality and pool variables under fencing or grazing conditions (n = 20).
Table 5.
The RMSE (g m−2) between simulated and observed nutritional quality and pool variables under fencing or grazing conditions (n = 20).
| Variables | Fencing | Grazing | ||||||
|---|---|---|---|---|---|---|---|---|
| Multiple Linear Regression | Random Forest | Support-Vector Machines | Recursive-Regression Trees | Multiple Linear Regression | Random Forest | Support-Vector Machines | Recursive-Regression Trees | |
| CP content | 0.60 | 0.13 | 0.32 | 0.40 | 0.66 | 0.19 | 0.53 | 0.55 |
| CP pool | 3.50 | 0.94 | 3.52 | 2.47 | 2.41 | 0.55 | 1.20 | 1.07 |
| Ash content | 3.47 | 0.86 | 2.18 | 2.47 | 3.80 | 0.46 | 2.39 | 2.49 |
| Ash pool | 7.58 | 2.37 | 7.67 | 8.31 | 10.24 | 2.99 | 6.68 | 5.97 |
| EE content | 0.30 | 0.12 | 0.16 | 0.17 | 0.33 | 0.07 | 0.15 | 0.22 |
| EE pool | 0.47 | 0.11 | 0.44 | 0.45 | 0.72 | 0.15 | 0.59 | 0.38 |
| WSC content | 0.84 | 0.32 | 0.79 | 0.61 | 1.61 | 0.33 | 1.04 | 0.92 |
| WSC pool | 1.22 | 0.32 | 0.89 | 0.84 | 2.88 | 0.50 | 1.61 | 1.46 |
| ADF content | 3.06 | 0.99 | 2.34 | 1.52 | 2.97 | 0.93 | 1.63 | 1.85 |
| ADF pool | 9.93 | 3.72 | 9.25 | 9.66 | 15.37 | 2.75 | 9.69 | 5.11 |
| NDF content | 2.43 | 0.61 | 2.01 | 1.47 | 6.80 | 0.74 | 3.77 | 2.26 |
| NDF pool | 14.45 | 4.50 | 7.88 | 5.83 | 17.55 | 2.80 | 11.09 | 5.93 |
Table 6.
Relative RMSE (%) between simulated and observed nutritional quality and pool variables under fencing or grazing conditions (n = 20).
Table 6.
Relative RMSE (%) between simulated and observed nutritional quality and pool variables under fencing or grazing conditions (n = 20).
| Variables | Fencing | Grazing | ||||||
|---|---|---|---|---|---|---|---|---|
| Multiple Linear Regression | Random Forest | Support-Vector Machines | Recursive-Regression Trees | Multiple Linear Regression | Random Forest | Support-Vector Machines | Recursive-Regression Trees | |
| CP content | 8.36 | 1.81 | 4.46 | 5.57 | 8.58 | 2.47 | 6.89 | 7.15 |
| CP pool | 84.07 | 22.58 | 84.55 | 59.33 | 67.16 | 15.33 | 33.44 | 29.82 |
| Ash content | 15.41 | 3.82 | 9.68 | 10.97 | 17.74 | 2.15 | 11.16 | 11.63 |
| Ash pool | 86.63 | 27.08 | 87.65 | 94.97 | 120.96 | 35.32 | 78.91 | 70.52 |
| EE content | 18.08 | 7.23 | 9.64 | 10.24 | 21.70 | 4.60 | 9.86 | 14.47 |
| EE pool | 65.91 | 15.43 | 61.71 | 63.11 | 121.12 | 25.23 | 99.25 | 63.92 |
| WSC content | 18.93 | 7.21 | 17.81 | 13.75 | 28.86 | 5.92 | 18.64 | 16.49 |
| WSC pool | 64.91 | 17.03 | 47.35 | 44.69 | 99.94 | 17.35 | 55.87 | 50.67 |
| ADF content | 7.55 | 2.44 | 5.77 | 3.75 | 7.42 | 2.32 | 4.07 | 4.62 |
| ADF pool | 49.77 | 18.65 | 46.36 | 48.42 | 71.36 | 12.77 | 44.99 | 23.72 |
| NDF content | 4.11 | 1.03 | 3.40 | 2.49 | 11.83 | 1.29 | 6.56 | 3.93 |
| NDF pool | 66.96 | 20.85 | 36.52 | 27.02 | 92.97 | 14.83 | 58.75 | 31.41 |
Table 7.
Relative bias (%) between simulated and observed nutritional quality and pool variables under fencing or grazing conditions (n = 20).
Table 7.
Relative bias (%) between simulated and observed nutritional quality and pool variables under fencing or grazing conditions (n = 20).
| Variables | Fencing | Grazing | ||||||
|---|---|---|---|---|---|---|---|---|
| Multiple Linear Regression | Random Forest | Support-Vector Machines | Recursive-Regression Trees | Multiple Linear Regression | Random Forest | Support-Vector Machines | Recursive-Regression Trees | |
| CP content | 4.16 | 0.71 | 1.48 | 1.33 | 0.63 | 0.90 | −0.20 | −0.82 |
| CP pool | −10.52 | 4.65 | −27.32 | −2.04 | −9.95 | 0.25 | −10.02 | 10.00 |
| Ash content | −2.11 | −0.78 | −1.46 | 0.02 | −4.57 | 0.91 | 4.31 | 2.02 |
| Ash pool | 18.86 | 3.44 | 11.62 | 23.37 | 45.75 | 4.38 | −11.59 | 22.45 |
| EE content | −1.79 | −0.96 | 1.71 | −0.75 | 13.16 | 1.23 | 1.79 | 6.14 |
| EE pool | 5.76 | −5.56 | −20.20 | 4.24 | 24.20 | 5.48 | 3.68 | 20.84 |
| WSC content | 0.23 | −1.78 | 0.52 | −5.16 | 1.10 | −0.18 | 1.13 | 1.19 |
| WSC pool | 4.48 | −1.02 | −7.99 | 3.55 | −6.05 | −1.44 | −18.71 | 9.19 |
| ADF content | −1.11 | −0.06 | −0.87 | −0.68 | −2.47 | −0.68 | −0.20 | −0.20 |
| ADF pool | 3.92 | 0.80 | −12.79 | 4.63 | −25.55 | −2.26 | −20.12 | −0.57 |
| NDF content | 0.28 | 0.51 | −0.30 | 0.16 | 1.62 | −0.32 | 1.43 | 0.48 |
| NDF pool | 21.04 | 0.83 | 3.71 | −2.23 | 28.43 | 2.65 | 5.52 | 19.89 |
Table 8.
The linear slopes between simulated and observed nutritional quality and pool variables under fencing or grazing conditions (n = 20).
Table 8.
The linear slopes between simulated and observed nutritional quality and pool variables under fencing or grazing conditions (n = 20).
| Variables | Fencing | Grazing | ||||||
|---|---|---|---|---|---|---|---|---|
| Multiple Linear Regression | Random Forest | Support-Vector Machines | Recursive-Regression Trees | Multiple Linear Regression | Random Forest | Support-Vector Machines | Recursive-Regression Trees | |
| CP content | 1.04 | 1.01 | 1.01 | 1.01 | 1.00 | 1.01 | 0.99 | 0.98 |
| CP pool | 0.55 | 1.02 | 0.47 | 0.77 | 0.71 | 0.99 | 0.90 | 1.01 |
| Ash content | 0.95 | 0.99 | 0.97 | 0.99 | 0.94 | 1.01 | 1.04 | 1.02 |
| Ash pool | 0.73 | 0.94 | 0.70 | 0.79 | 0.84 | 0.98 | 0.59 | 1.30 |
| EE content | 0.94 | 0.99 | 1.00 | 0.98 | 1.12 | 1.01 | 1.01 | 1.05 |
| EE pool | 0.75 | 0.95 | 0.60 | 0.86 | 0.80 | 1.00 | 0.64 | 0.92 |
| WSC content | 0.97 | 0.98 | 0.97 | 0.94 | 0.94 | 0.99 | 0.97 | 0.99 |
| WSC pool | 0.77 | 0.97 | 0.77 | 0.91 | 0.72 | 1.00 | 0.70 | 1.00 |
| ADF content | 0.99 | 1.00 | 0.99 | 0.99 | 0.97 | 0.99 | 1.00 | 1.00 |
| ADF pool | 0.81 | 0.97 | 0.73 | 0.92 | 0.62 | 0.98 | 0.76 | 0.95 |
| NDF content | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
| NDF pool | 0.94 | 0.99 | 0.90 | 0.94 | 0.91 | 1.00 | 0.91 | 1.01 |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
MDPI and ACS Style
Han, F.; Fu, G.; Yu, C.; Wang, S. Modeling Nutrition Quality and Storage of Forage Using Climate Data and Normalized-Difference Vegetation Index in Alpine Grasslands. Remote Sens. 2022, 14, 3410. https://doi.org/10.3390/rs14143410
AMA Style
Han F, Fu G, Yu C, Wang S. Modeling Nutrition Quality and Storage of Forage Using Climate Data and Normalized-Difference Vegetation Index in Alpine Grasslands. Remote Sensing. 2022; 14(14):3410. https://doi.org/10.3390/rs14143410
Chicago/Turabian StyleHan, Fusong, Gang Fu, Chengqun Yu, and Shaohua Wang. 2022. "Modeling Nutrition Quality and Storage of Forage Using Climate Data and Normalized-Difference Vegetation Index in Alpine Grasslands" Remote Sensing 14, no. 14: 3410. https://doi.org/10.3390/rs14143410
Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.
