Carbon Cycles of Forest Ecosystems in a Typical Climate Transition Zone under Future Climate Change: A Case Study of Shaanxi Province, China
Round 1
Reviewer 1 Report
The manuscript examined the modeled response of forest NPP, NEP, and Rh to climate change in a typical climate transition zone in China. Different climate scenarios and two types of forests were investigated. I recommend a major revision along with the following comments and questions.
I suggest the authors summarize the future climate projections in terms of changes in temperature, precipitation, and CO2, etc. Therefore, readers can have a picture of the future trend for the study area. The authors need to address more the difference between Q. wutaishanica and R. pseudoacacia forests as well as the purpose of simulating them separately. They both fall in the functional deciduous broadleaf species. DGVM modeled normally consider them as functionally similar. One big concern about the methods is that the model was spun up without any anthropogenic disturbances. As the text mentioned the forests are plantations and secondary forests, and the areas are part of the ‘Grain for Green’ program; prior agricultural disturbance should be an important factor for the reforestation. Many studies show the legacy of prior disturbance can last hundreds of years. The studied forests look young with small DBH and short hight (Table S1). On the other hand, after spinup a forest normally reaches its equilibrium stage, implying an old-growth. This discrepancy probably resulted in the underestimation of NPP (Fig. 2) by the model and relative low predicted NEP (Figs. 3 and 4) and the fact that NEP quickly approaches low values under future scenarios. Cabon cycle is closely affected by the N cycle. The change in N deposition is a very important factor of global change altering functions of ecosystems. Prior disturbance can have a huge legacy on the N status of forests and, therefore carbon cycle. The study should address or discuss the uncertainty of the effect of N cycling on the carbon cycle, e.g., N deposition. At present, the study is only focused on temperature, precipitation, and atmospheric CO2 concentration. The validation seems weak in this study. I expect to see the yearly NPP comparison between measurements and predictions. Are there any eddy covariance data for NEP validation? Table 1 is not clear. What is observed and what is simulated? Why did the study only examine the sensitivity of NPP instead of NPP and NEP? The discussion about the future trends of NEP and Rh are lack of details. Why does the NEP have a declining trend in the future Rh will reach a saturated level?Author Response
Reviewer 1:
I suggest the authors summarize the future climate projections in terms of changes in temperature, precipitation, and CO2, etc. Therefore, readers can have a picture of the future trend for the study area.
Response: Thank you very much for your suggestion. We have revised the content related to this concern on Pages 1−2, Lines 34−36 and 42−53.
“Since the industrial revolution, atmospheric CO2 concentration has increased from 280 ppmv before the industrial revolution to 400 ppmv at present, and the global average surface temperature has increased by 0.85 °C.”
“In China, since the 1990s, the average temperature has increased at the rate of 0.6 °C 10 yr-1, which is far higher than the global average of 0.27 °C 10 yr-1, and together with average temperature, seasonal precipitation has also changed significantly. These climate changes are particularly evident in northern China. The Shaanxi Province, located in the north of China, is one of the regions where climate change is obvious. Our previous study predicted that the temperature in this area would rise 1.4−3.1 °C by the end of this century, while precipitation showed no obvious change. This region is also a transition zone from a temperate arid to a subtropical humid climate ”
The authors need to address more the difference between Q. wutaishanica and R. pseudoacacia forests as well as the purpose of simulating them separately. They both fall in the functional deciduous broadleaf species.
Response: We have provided this information in the revised version of the manuscript, Page 3, Lines 112−121.
“Robinia pseudoacacia has become the main tree species in afforestation and has the largest plantation area in Shaanxi Province because of its developed root system, rapid growth, drought tolerance, barren tolerance, and high survival rate. Quercus L. (Quercus wutaishanica being the dominant species) has a well-developed root system, cold and drought tolerance, strong adaptability, and a wide distribution area in Shaanxi Province, being an important natural vegetation type in this area. R. pseudoacacia and Q. wutaishanica are the representative tree species of artificial forest and natural secondary forest in Shaanxi Province, accounting for 30.7% and 63.0%, respectively, of the carbon reserves in plantations and natural secondary forests in Shaanxi Province. Small changes in their carbon sequestration may cause large carbon fluctuations in the Shaanxi Province.”
DGVM modeled normally consider them as functionally similar. One big concern about the methods is that the model was spun up without any anthropogenic disturbances. As the text mentioned the forests are plantations and secondary forests, and the areas are part of the ‘Grain for Green’ program; prior agricultural disturbance should be an important factor for the reforestation. Many studies show the legacy of prior disturbance can last hundreds of years. The studied forests look young with small DBH and short hight (Table S1). On the other hand, after spinup a forest normally reaches its equilibrium stage, implying an old-growth. This discrepancy probably resulted in the underestimation of NPP (Fig. 2) by the model and relative low predicted NEP (Figs. 3 and 4) and the fact that NEP quickly approaches low values under future scenarios.
Response: This passage in the introduction may have caused a misunderstanding. In fact, the sample plots we selected are not interfered by humans or the interference is weak and can be ignored. Therefore, we deleted this sentence to avoid creating confusion.
Shaanxi Province is an area with serious soil erosion and desertification in China. The poor growth environment makes forest growth in this area different from that in other areas. According to your suggestion, we have re-filtered the samples collected in the field, and raised the standard of tree DBH from 5 to 8 cm. On this basis, we have verified the model and Figure 2 is the result of this revalidation. For NEP, only the forest in NSX area has low NEP, and is likely to become a carbon source. We believe that this is mainly the result of higher temperature and less precipitation in NSX area under future climate change, while MSX and SSX areas with suitable water and heat, and high forest NEP, are a larger carbon sink in the future climate change scenario.
Cabon cycle is closely affected by the N cycle. The change in N deposition is a very important factor of global change altering functions of ecosystems. Prior disturbance can have a huge legacy on the N status of forests and, therefore carbon cycle. The study should address or discuss the uncertainty of the effect of N cycling on the carbon cycle, e.g., N deposition.
Response: The plots we selected are not interfered by humans or the interference is weak and can be ignored. In addition, in this study, we used LPJ-GUESS version 3.0, which is an updated version including the interaction between atmospheric nitrogen cycle and carbon cycle. In order to explain the model more accurately, we introduced the model further on Page 4, Lines 129−132, and discussed it on Page 13, Lines 381−388 and 396−402.
“The simulation process is carried out in plant functional types (PFTs) or species, Sitch et al. described the simulation scheme of LPJ-GUESS model. In this study, we used model version 3.0, which is an updated version including the interaction between atmospheric nitrogen cycle and carbon cycle, completely described by Smith et al. ”
“Additionally, studies have shown that atmospheric nitrogen deposition can significantly reduce soil Rh in temperate forests, which is attributed to the decrease of soil microbial biomass and litter decomposition rate after high level of nitrogen deposition. Contrastingly, other studies have come to the opposite conclusion, which is attributed to the increase of soil microbial biomass and activity, followed by the increase of soil Rh. In this study, the increase of soil Rh with increasing climate emission intensity is not caused by one factor, but the result of multiple factors. As the study did not investigate nitrogen deposition, whether nitrogen deposition was positive or negative is uncertain, which would be the direction of our future research.”
“In this study, our simulation results showed that, in the arid area, forest NEP would decrease, and forests may become a carbon source in the future. This may be mainly attributed to the lower soil moisture content in the arid area, as potential evapotranspiration in this region is greater than precipitation. Contrastingly, in the humid area, forest NEP will increase, and forests will become a larger carbon sink in the future, which is attributed to the water and heat state in this area being more suitable for forest growth, as well as the effect of atmospheric nitrogen deposition being more apparent in humid regions.”
At present, the study is only focused on temperature, precipitation, and atmospheric CO2 The validation seems weak in this study.
Response: Our research dealt with the response of forest carbon cycle to climate change, and temperature, precipitation and atmospheric CO2 concentration are the three most important aspects of climate change.
I expect to see the yearly NPP comparison between measurements and predictions.
Response: This is a great suggestion. We have modified Figure 2 to accommodate this suggestion. Please refer to figure 2 of the revised manuscript.
Are there any eddy covariance data for NEP validation? Table 1 is not clear. What is observed and what is simulated?
Response: We don't have eddy covariance data. Table 1 represents the comparison between simulated and field-based NEP values from the references (Page 5 Lines 197−199). It is widely known that NEP is different from NPP and Rh, its continuity is poor and the interannual fluctuation is large. Therefore, when the field-based NEP value is between the simulated values and there is no significant difference between the simulated and the field-based NEP value, we assumed that the model simulation was good. In addition, we have revised table 1. Please refer to the revised manuscript (Lines 212−213).
Why did the study only examine the sensitivity of NPP instead of NPP and NEP?
Response: We have explained in the introduction that our purpose was to study the sensitivity of Robinia pseudoacacia and Quercus wutaishanica forests to temperature, precipitation, and atmospheric CO2 concentration in Shaanxi Province, China, and NPP is only used as a research method. As explained in the revised manuscript, previous studies have shown that NPP is more sensitive to climate change and is suitable for measuring the sensitivity of forest ecosystems to climate change. Therefore, NPP was selected as an indicator of forest sensitivity in this study (Page 2 Lines 67−70).
The discussion about the future trends of NEP and Rh are lack of details. Why does the NEP have a declining trend in the future Rh will reach a saturated level?
Response: Thank you very much for your suggestion. We may have misrepresented our conclusion, forest NEP in the arid region will decrease while that in the humid region will increase. In addition, forest soil Rh in both arid and humid regions will increase, rather than reaching saturation level. However, we did ignore some factors that affect Rh and NEP of forests, which we have mentioned in the revised discussion section. Please refer to Lines 381−388 and 396−402 on Page 13.
“Additionally, studies have shown that atmospheric nitrogen deposition can significantly reduce soil Rh in temperate forests, which is attributed to the decrease of soil microbial biomass and litter decomposition rate after high level of nitrogen deposition. Contrastingly, other studies have come to the opposite conclusion, which is attributed to the increase of soil microbial biomass and activity, followed by the increase of soil Rh. In this study, the increase of soil Rh with increasing climate emission intensity is not caused by one factor, but the result of multiple factors. As the study did not investigate nitrogen deposition, whether nitrogen deposition was positive or negative is uncertain, which would be the direction of our future research.”
“In this study, our simulation results showed that, in the arid area, forest NEP would decrease, and forests may become a carbon source in the future. This may be mainly attributed to the lower soil moisture content in the arid area, as potential evapotranspiration in this region is greater than precipitation. Contrastingly, in the humid area, forest NEP will increase, and forests will become a larger carbon sink in the future, which is attributed to the water and heat state in this area being more suitable for forest growth, as well as the effect of atmospheric nitrogen deposition being more apparent in humid regions.”
Author Response File: 
Author Response.doc
Reviewer 2 Report
Liang et al. show NPP and Rh simulations based on LPJ-GUESS model in the Shaanxi Province of China. The topic is quite interesting and suitable for Forests. However, I think the manuscript in current version has several major issues, thereby this manuscript needs major revision before I recommend it to be accepted for publication in Forests.
Model validationAuthors only show annual mean NPP based on LPJ-GUESS simulation as compare to measured (observed) NPP in Figure 2. I suggest that the annual NPP cycle and interannual variation of NPP should be present here compared to observation. Authors tried to understand future changes of NPP in terms of mean and trend. Even though LPJ-GUESS shows reasonable mean NPP value in the historical period in Figure 2, there is no guarantee for how this model well represents future changes. If interannual variation of NPP is reasonable as a result of the response to temperature and precipitation, we can conclude the confidence of future trend in NPP changes. I strongly recommend providing validation of interannual variation of NPP in the historical period. (2001 to 2010?)
NPP = GPP - Ra
I am not sure that observational data cannot provide GPP and Ra separately, but model results would have GPP and Ra to calculate NPP. It might be great to discuss which one is a dominating factor of driving future change of NPP (GPP or Ra)
Variation changes
Authors only described average value and trend in Table 2 and 3, but it seems variations are also amplified in Figure 3 to 6. As you mentioned in several parts of the manuscript, this information could be applied to help policy makers in planning sustainable forests management. How about providing standard deviation in Table 2 and 3? You can also do same calculations in Table 4 and 5.
Schemes in LPJ-GUESS
You mentioned that root respiration scheme in the section 4.4, but you missed soil respiration, photosynthesis, stomatal conductance and many others in your manuscript. Especially, soil respiration shows temperature-related future changes are important, so you need to descript which scheme used in LPJ-GUESS such as Q10 value. On the other hand, the stomatal conductance scheme (Ball-Berry?) can also contribute to the future change of NPP. Change of sensitivities
Some of previous studies argued that sensitivity of forests to temperature and precipitation would be changed as following changing of mean climate status (Liu et al. 2016 and Kim et al. 2017). It is needed to describe previous studies more carefully.
Liu, Y. W. et al. Changes in interannual climate sensitivities of terrestrial carbon fluxes during the 21st century predicted by CMIP5 earth system models. J. Geophys. Res. 121, 903–918 (2016).
Kim, J.-S. et al. Intensification of terrestrial carbon cycle related to El Niño–Southern Oscillation under greenhouse warming, Nat. Comm. 8, 1674 (2017).
Minor concerns
L237 You did model experiments, but I think it is (very) ideal cases. Indeed, when the temperature increased, precipitation is also changed in real climate system. Also, while ideal model experiments give some insights, but 30% changes of temperature and 30% of precipitation have a different physical meaning and also quantitively. 30% is based on anomaly value? It is needed to describe including more detail.
L281 Do you have evapotranspiration results from LPJ-GUESS?
Figure 1 Can you write where is NSX, MSX and SSX in Figure 1?
Author Response
Reviewer 2:
Model validation
Authors only show annual mean NPP based on LPJ-GUESS simulation as compare to measured (observed) NPP in Figure 2. I suggest that the annual NPP cycle and interannual variation of NPP should be present here compared to observation. Authors tried to understand future changes of NPP in terms of mean and trend. Even though LPJ-GUESS shows reasonable mean NPP value in the historical period in Figure 2, there is no guarantee for how this model well represents future changes. If interannual variation of NPP is reasonable as a result of the response to temperature and precipitation, we can conclude the confidence of future trend in NPP changes. I strongly recommend providing validation of interannual variation of NPP in the historical period. (2001 to 2010?)
Response: Thank you very much. In fact, this is a great suggestion. We have modified Figure 2 in the revised manuscript.
NPP = GPP - Ra
I am not sure that observational data cannot provide GPP and Ra separately, but model results would have GPP and Ra to calculate NPP. It might be great to discuss which one is a dominating factor of driving future change of NPP (GPP or Ra)
Response: Unfortunately, we don't have the observation data of GPP and Ra, and although NPP in the model is obtained by subtracting Ra from GPP, we can verify the model well by comparing the simulation and field value of NPP. Therefore, although the reviewer's suggestion is interesting, it is out of the scope of the manuscript, as we did not focus on the particular trends of NPP under future climate change in our manuscript, but analyzed the sensitivity of forest to climate factors by simulating different climate scenarios. In addition, temperature, precipitation and CO2 are the three most important variables of climate change in the future, and they are all the leading factors to promote the change of NEP and Rh. But the effects of the three variables on arid and humid areas are slightly different, which we have also discussed in the discussion section.
Variation changes
Authors only described average value and trend in Table 2 and 3, but it seems variations are also amplified in Figure 3 to 6. As you mentioned in several parts of the manuscript, this information could be applied to help policy makers in planning sustainable forests management. How about providing standard deviation in Table 2 and 3? You can also do same calculations in Table 4 and 5.
Response: Thank you very much for your advice. We have added standard deviations to tables 2 and 3 in the manuscript.
Schemes in LPJ-GUESS
You mentioned that root respiration scheme in the section 4.4, but you missed soil respiration, photosynthesis, stomatal conductance and many others in your manuscript. Especially, soil respiration shows temperature-related future changes are important, so you need to descript which scheme used in LPJ-GUESS such as Q10 value. On the other hand, the stomatal conductance scheme (Ball-Berry?) can also contribute to the future change of NPP.
Response: There will be many uncertainties in the ecosystem model. In fact, we discussed the root respiration scheme in section 4.4 just to explain the limitations of the ecosystem model itself, which we cannot solve. Therefore, we propose this problem in the uncertainty analysis. In addition, we did ignore the simulation scheme of the model in the discussion because there are many simulation schemes in the model, and we can't describe them all in the manuscript, so we made the following modifications on Pages 3−4, Lines 127−132.
"LPJ-GUESS is a multi-scale dynamic process model for simulating ecosystem structure and function; it can perform simulations on a variety of scales, including patches, landscapes, regions, and even global scale simulations. The simulation process is carried out in plant functional types (PFTs) or species, Sitch et al. described the simulation scheme of LPJ-GUESS model. In this study, we used model version 3.0, which is an updated version including the interaction between atmospheric nitrogen cycle and carbon cycle, completely described by Smith et al. "
Furthermore, we didn't focus on the temperature sensitivity of heterotrophic respiration. We just used the model to simulate the trend of forest Rh change under the future climate change scenarios (not only the temperature change), and analyzed the possible reasons for this change. Therefore, Q10 was not included in our simulation scheme. Finally, we agree that stomatal conductance is an indirect factor affecting forest NPP, which is also affected by climate change, and we have taken this factor into account in the discussion (Section 4.1 in the revised manuscript).
Change of sensitivities
Some of previous studies argued that sensitivity of forests to temperature and precipitation would be changed as following changing of mean climate status (Liu et al. 2016 and Kim et al. 2017). It is needed to describe previous studies more carefully.
Liu, Y. W. et al. Changes in interannual climate sensitivities of terrestrial carbon fluxes during the 21st century predicted by CMIP5 earth system models. J. Geophys. Res. 121, 903–918 (2016).
Kim, J.-S. et al. Intensification of terrestrial carbon cycle related to El Niño–Southern Oscillation under greenhouse warming, Nat. Comm. 8, 1674 (2017).
Response: Thank you very much for your suggestion. We have revised the manuscript accordingly. Pages 11–12, Lines 291-299, 327-328, 334-344, and 351-357.
“Our research showed that increased precipitation had a great positive effect on NPP in the arid region, which is attributed to increased precipitation supplementing soil water content, reducing water stress caused by excessive transpiration, and increasing stomatal conductance, thus increasing forest CO2 intake and NPP. Forest growth was also observed to be more sensitive to precipitation change in arid and semi-arid areas such as the south of the United States, Central Asia, the south of South America, the south of Africa, and Australia, and in temperate semi-arid areas of the southern hemisphere than in humid areas, which is attributed to the precipitation increase reducing NPP by decreasing radiation input, increasing nutrient leaching, or reducing the soil oxygen availability when soil moisture can already support forest growth.”
“Considering solely the temperature increase, forest NPP decreased in both arid and humid regions. ”
“However, this conclusion may not be suitable for forest ecosystems in middle and high latitudes and high altitudes. Previous studies have shown that the sensitivity of forest NPP to temperature is higher with the increase of future temperature in the northwest of the United States, and forest NPP increases with increasing temperature in the Tianshan and Qilian Mountains of China. This difference should be attributed to the fact that temperature is the main factor limiting forest growth in the middle and high latitudes and high altitudes. Temperature rise can promote forest photosynthesis in these areas, which is one of the important processes in NPP growth. Photosynthesis response to temperature can be described as a parabola around the best temperature value. Below this temperature, photosynthesis increases with the increase of temperature, and above this temperature, photosynthesis decreases with the increase of temperature owing to the decrease of stomatal conductance and transpiration.”
“This result is different from previous studies. One study found that forest NPP increased by only about 2.7% under increased CO2 concentration alone (from 355 to 710 ppmv), and another study found that forest NPP declined under a 12-year sustained CO2 fertilization experiment (maintained at 550 ppmv). These differences may be attributed to the fact that the CO2 concentration used in this study did not reach the critical point of intracellular carboxylation reaction, i.e., we increased CO2 concentration by a maximum of 30% on the basis of the historical period, to about 494 ppmv. ”
Minor concerns
L237 You did model experiments, but I think it is (very) ideal cases. Indeed, when the temperature increased, precipitation is also changed in real climate system. Also, while ideal model experiments give some insights, but 30% changes of temperature and 30% of precipitation have a different physical meaning and also quantitively. 30% is based on anomaly value? It is needed to describe including more detail.
Response: We admit that the model experiment we have done is very ideal, but the purpose of our experiment is also very simple, just to know if forests in arid and humid areas have different sensitivities to climate factor changes in the future scenario. For this purpose, we did the model experiment. Our previous study concluded that the maximum possible increase of future temperature in the study area is about 30%, and the average atmospheric CO2 concentration (510 ppmv) in the future is about 30% higher than the current atmospheric CO2 concentration (380 ppmv). In order to measure uniformly, we set the maximum increase of the three indicators of sensitivity analysis as 30%, and compared them with the current value (0%) and the intermediate value (15%). We have explained this in the revised manuscript. Please refer to Line 169−174 on Page 4.
“Our previous study concluded that the maximum possible future increase of temperature in the study area is about 30%, and the average atmospheric CO2 concentration in the future (510 ppmv) will be about 30% higher than the current atmospheric CO2 concentration (380 ppmv). To measure the variables uniformly, we set the maximum increase of the three indicators of sensitivity analysis as 30%, and compared them with the current value (0%) and the intermediate value (15%).”
L281 Do you have evapotranspiration results from LPJ-GUESS?
Response: Yes, we have. Please refer to Peng et al. (2018) - Spatiotemporal change and trend analysis of potential evapotranspiration over the Loess Plateau of China during 2011−2100. Cited in the revised manuscript as reference [23]
Figure 1 Can you write where is NSX, MSX and SSX in Figure 1?
Response: Thank you very much for pointing this out. It was indeed our negligence. We have corrected Figure 1.
Author Response File: Author Response.doc
Round 2
Reviewer 2 Report
After major revision, I found the improved manuscript and it is worth to publish in Forests. There are only a few minor concerns before publishing.
Citing Kim et al. 2017
L289-291 Forest growth was also observed to be more sensitive to precipitation change in arid and semi-arid areas such as the south of the United States, Central Asia, the south of South America, the south of Africa, and Australia [27,28]
28: Kim, J.-S. et al. Intensification of terrestrial carbon cycle related to El Niño–Southern Oscillation under greenhouse warming, Nat. Comm. 8, 1674 (2017).
Minus symbol
Minus symbol should be edited from - to − in Table 4 and 5.
(https://en.wiktionary.org/wiki/minus_sign)
P4 L141 Representative Concentration Pathway (RCP)
(You should mention the full name of RCP here, before L153-154)
Capitalization for NEP, NPP and Rh
L16, L19, L52, L56 and so on
Net Primary Productivity (NPP)
Net Ecosystem Productivity (NEP)
Heterotrophic Respiration (Rh)
Author Response
Citing Kim et al. 2017
L289-291 Forest growth was also observed to be more sensitive to precipitation change in arid and semi-arid areas such as the south of the United States, Central Asia, the south of South America, the south of Africa, and Australia [27,28]
28: Kim, J.-S. et al. Intensification of terrestrial carbon cycle related to El Niño–Southern Oscillation under greenhouse warming, Nat. Comm. 8, 1674 (2017).
Minus symbol
Minus symbol should be edited from - to − in Table 4 and 5.
(https://en.wiktionary.org/wiki/minus_sign)
P4 L141 Representative Concentration Pathway (RCP)
(You should mention the full name of RCP here, before L153-154)
Capitalization for NEP, NPP and Rh
L16, L19, L52, L56 and so on
Net Primary Productivity (NPP)
Net Ecosystem Productivity (NEP)
Heterotrophic Respiration (Rh)
Response：Thank you very much for your advice, we have revised it in the manuscript. In fact, this is a very good suggestion. The problem of format can reflect the seriousness and working attitude of an author. To be honest, we haven't done well enough in this respect. Thank you very much for asking these details, which not only makes the article better, but also makes us aware of the importance of details, so that we can do better next time we submit the manuscript.
