Can the Pinus sylvestris var . mongolica sand-fixing forest develop sustainably in a semi-arid region ?

Desertification is a global environmental and societal concern at present, and China is one of the countries 10 that face the most severe damage of desertification. China’s so-called Three North shelterbelt Program (3NSP) has produced a vast area of lined forest in the semi-arid regions with the purpose of battling desertification. Such a windbreaking and sand-fixing forest has successfully slowed down the incursion of desert. However, the vast artificial forestry consumes a large amount of water resources, which profoundly affect the fragile ecological environment in the semi-arid regions. In turn, a large amount of water loss also causes a great number of vegetation deaths or defects. 15 To understand the water balance and sustainable development of artificial forest in semi-arid region, this study uses the 30-year-old lined Pinus sylvestris var. mongolica sand-fixing forest in the eastern part of Mu Us Sandy land in Northwestern China as an example. Specifically, this investigation studies the redistribution of water in soil under existing precipitation conditions, so as to evaluate whether the rain-feed forestry can develop sustainably or not. Rain gauge, newly designed lysimeter and soil moisture sensor are used to monitor precipitation, deep soil recharge (DSR) 20 and soil water content, resulting in an accurate estimation of annual moisture distribution of the rain-feed Pinus sylvestris var. mongolica. The study shows that there are two obvious moisture recharge processes in an annual base for the Pinus sylvestris var. mongolica forest soil in Mu Us Sandy land: 1) the snow melted water infiltration-recharge process in the spring, and 2) the precipitation-recharge process in the summer. The recharge depth of the first process is 160 cm. The second process results in DSR (referring to recharge that can reach a depth more than 200 cm and may 25 eventually replenish the groundwater reservoir). The DSR of 2016-2018 is 1.4 mm, 0.2mm, 1.2 mm, respectively. To reach the recharge depths of 20 cm, 40 cm, 80 cm, 120 cm, 160 cm, and 200 cm, the corresponding precipitation intensities have to be 2.6 mm/d, 3.2 mm/d, 3.4 mm/d, 8.2 mm/d, 8.2 mm/d, and 13.2 mm/d, respectively. The annual evaporation amount in the Mu Us Sandyland Pinus sylvestris var. mongolica forest is 426.96 mm in 2016, 324.6 mm in 2017, 416.253 mm in 2018. This study concludes that under the current precipitation conditions, very small but 30 observable DSR happened, thus the groundwater system underneath the forest may be replenished, meaning that the artificial Pinus forestry can probably develop sustainably. This study confirms that developing limited amount forestry in semi-arid regions is likely in a sustainable fashion. The widely variable annual precipitation in semi-arid areas may affect this conclusion and should be investigated in the future. Hydrol. Earth Syst. Sci. Discuss., https://doi.org/10.5194/hess-2019-110 Manuscript under review for journal Hydrol. Earth Syst. Sci. Discussion started: 18 March 2019 c © Author(s) 2019. CC BY 4.0 License.


Introduction
Desertification is a global threat that impacts heavily on the livelihoods of millions of people inside and outside the desert land (Yang et al., 2005).The true cost of desertification is frequently underestimated due to the unknown scale of these external and downstream impacts (Cao et al., 2011).Sand and dust storms happen when strong winds impact the arid and semi-arid regions (Sun et al., 2015;Wang et al., 2010a).Sandstorms occur relatively close to the Hydrol.Earth Syst.Sci.Discuss., https://doi.org/10.5194/hess-2019-110Manuscript under review for journal Hydrol.Earth Syst.Sci. Discussion started: 18 March 2019 c Author(s) 2019.CC BY 4.0 License.a sizable DSR can occur, meaning that groundwater may be recharged from precipitation, the sustainable reforestation in the region is possible.To accomplish this goal, this study uses a newly developed DSR lysimeter to monitor a 30-105 year old pine artificial forest in the Northwest China.The collected dataset is used to understand the soil moisture dynamic and the DSR of the Pinus sylvestris var.mongolica in sandy land.Specifically, we try to tackle the following issues: 1) Sources of soil water recharge in semi-arid areas, especially the source of deep soil layer moisture; 2) The precipitation density that causes infiltration and its maximal penetrating depth; 3) The rate of annual precipitation infiltration; 4) The evaporation amount of the pine forestry land.The ultimate goal is to find out whether the rain-feed 110 Pinus sylvestris var.mongolica sand-fixing forest can develop sustainably or not in the study site.

Overview of the study area
The area under study is located in Chagan Naoer, on the northeastern edge of Mu Us Sandy land (39°05′16.2″N,109°36′04″E), as shown in Figure 1.Mu Us Sandy land mainly consists of semi-fixed and fixed sand dunes, adjacent 115 with the Loess Plateau, located in a desert-loess transitional zone.It has northwestern wind in the winter with a typically dry winter climate and frequent sandstorm.It has southeastern monsoon in the summer.The summer climate is relatively humid, and it is easy to form local heavy precipitation.The multi-year average precipitation is 400 mm, mostly concentrated during the summer.The groundwater table depth varies between 2 m to17 m in Mu Us Sandyland, and it is 8 m at the experimental area of this study (Runnström, 2003).The groundwater table is lower in the summer 120 and higher in the spring, with a variation less than 1.5 m.Since the initiation of 3NSP at Mu Us Sandy land in 1989, Pinus sylvestris var.mongolica has been planted in lines in the experimental area, with a 10 m line spacing, an average plant height of 6 m, and an average crown diameter of 6.6 m.The seasonal frozen soil period in the experimental area is from January to April, and from November to December in an annual base (Li et al., 2013a).

Experimental design
Root analysis of Pinus sylvestris var.mongolica shows that it has a shallow root distribution and underdeveloped main root, which belongs to a typical lateral root type.The 30-year-old Pinus sylvestris var.mongolica's root distribution can reach 6.5 m, with a concentrated area of 2.5 m.In this study site, the ground water level is too deep 130 to supply for roots, so Pinus sylvestris var.mongolica root is mostly distributed over a vertical range of 0-0.5m and relies on precipitation for water supply.The original main root usually stops growing at 1.5-2 m depth (Zhu et al., 2006).Therefore, in the experimental design of this study, the lowest soil moisture sensor is placed at a depth of 200 cm, and the lysimeter placement is also at the 200 cm depth.The canopy of Pinus sylvestris var.mongolica is capable of intercepting precipitation, thus affects the measurements of precipitation and soil moisture directly underneath the 135 canopy (Roth et al., 2007).Therefore, the measurements of this study are made in the middle between the forest lines, without the interference of canopy.

Soil moisture monitoring
Based on the root depth of Pinus sylvestris var.mongolica, the depth range of soil moisture sensor placement is determined.A soil section is cut out in the middle between two forest lines.The section consists of a layer of dead 140 tree leaves, a leached layer, a depositional layer, and a native soil layer, which is of fine sand.Soil moisture sensors (MiniTRASE TDR, USA) are placed in soil layers at 20 cm, 40 cm, 80 cm, 120 cm, 160 cm, and 200 cm depths.The measurement interval is one hour.

DSR monitoring
To study the moisture distribution of Pinus sylvestris var.mongolica in Mu Us Sandy land, two sets of data need 145 to be collected: precipitation from a rain gauge, and DSR measurement from a lysimeter.Surface runoff does not exist in the experimental area, thus is not a concern.Precipitation is monitored by rain gauge (Spectrum, USA, accuracy 0.2 mm), placed 1.5 m above ground surface.This study uses a new lysimeter to measure the DSR (Cheng et al., 2017).
Such a new lysimeter has two parts: an upper water balance part and a lower measurement part.As shown in Figure 2, the water balance part uses a cylindrical impermeable side wall to enclose a soil column for measurement.The 150 length of soil column is determined based by the capillary rise, which is approximately 60 cm, based on the soil particle size distribution at the site.The advantage of this design is that when soil at point B in Figure 2 reaches saturation, the capillary water reaches point A. Therefore, when infiltrated water enters the water balance part at depth A, additional infiltrated water after satisfying the saturation of soil between A and B will go into the measuring part.
The measurement part has a measurement accuracy of 2 mm (Spectrum, USA).The lysimeter is placed at 200 cm 155 depth to measure DSR, meaning that any precipitation-induced infiltration passing the 200 cm depth will not be subject to evaportranspirative process anymore.In another word, the infiltrated water that can pass the 200 cm depth of soil will keep going down and may become groundwater recharge.Before taking the measurements, the new lysimeter needs to be placed one year in advance, going through naturally settlement for a year, and allowing the soil to reach its pre-installation condition.

Results and Discussion
The soil moisture variation of Pinus sylvestris var.mongolica in 2016 is shown in Figure 3.It reveals that soil 165 moisture has obvious seasonal variational trends.The soil from January to March is frozen.The near surface soil moisture recharge is from snowmelt.When the near surface frozen soil starts to thaw, soil at the 20 cm depth is recharged on February 9 th , 16 th and 26 th in 2016.Soil at depths greater than 20 cm remains relatively stable.Frequent precipitation events usually occur from June to November, during which soil moisture changes considerably, and soil moistures at different depths exhibit periodic increase or decrease, regulated by the interplay of precipitation and 170 evapotranspiration.After February 26 th in 2016, soil gradually thaws completely.Figure 3 shows that snowmelt can recharge the soil moisture as deep as 140 cm.The soil moisture at 200 cm depth is recharged for the first time after a heavy precipitation event on July 8 th in 2016.

175
In order to study the degree of soil moisture response to precipitation in individual layers, this research choose each layer's soil moisture at the beginning of each month as a representative, to observe whether the soil moisture in a specific layer is recharged.Figure 4 shows the soil moistures at depths of 20 cm, 40 cm, 80 cm, 120 cm, 160 cm, and 200 cm at the beginning of each month.From Figure 3, the soil of Pinus sylvestris var.mongolica exhibits four distinctive layers: an evaporation layer at 0-40 cm depth, a lateral root activity layer at 40-160 cm depth, a dry soil 180 layer at 160-200 cm depth, and a deep soil layer below 200 cm.For the 0-40 cm evaporation layer, the soil moisture increases only under the effect of precipitation or snowmelt.Its moisture content decreases rapidly under the interplay of evaporation and infiltration.For the 40-160 cm root activity layer, the soil moisture is recharged from infiltrated water passing through the upper layer, and it gradually decreases under the effects of infiltration and root moisture absorption.For the 160-200 cm dry soil, the infiltrated water hardly reaches this layer, and the soil layer with a soil 185 moisture under the withered point is formed.The deep soil below 200 cm depth is of native sand soil, and Figure 3 shows that the soil moisture content of this layer is recharged four times under heavy precipitation events in 2016.As shown in Figure 6, the layers of strong soil moisture fluctuations are 20 cm, 40 cm, 80 cm, 120 cm, and 160 cm.The 200 cm level of soil changes relatively small, and was only recharged on September 7 th of 2018.The annual 215 soil moisture infiltration is only 1.2 mm in 2018, which is 1 mm higher than that in 2017 and 0.2 mm lower than that in 2016.

Recharge intensity of different soil layer infiltration
In the laboratory, one may calculate the precipitation intensity infiltrating into a specific soil layer according to 245 the soil characteristics with a proper mathematical model.In the natural environment, however, there are too many factors affecting the infiltration process, such as temperature and air humidity, wind speed, surface soil moisture, soil heterogeneity, etc.In this study, we computed the replenishment of each soil layer by analyzing the precipitationinduced wetting point to find out the minimum precipitation intensity for each individual layer.
Precipitation infiltration results in increasing soil moisture content.Each time when soil moisture infiltrates into  1, which shows that when precipitation infiltrates to soil layers at 20 cm, 40 cm, 80 cm, 120 cm, 160 cm, and 200 cm depths, the minimum precipitation intensities are 2.6 mm/d, 3.2 mm/d, 3.4 mm/d, 8.2 mm/d, 8.2 mm/d, 255 and 13.2 mm/d, respectively.When heavy precipitation events (precipitation intensity higher than 100 mm/d) happen in the study site, precipitation can recharge the soil moisture at 200 cm depth.
where P is precipitation, ET is evapotranspiration, and δW is the whole 200 cm soil layer moisture change.Runoff is not included in above water balance equation because it does not occur during the experiment.As P, DSR, and δW can be accurately measured, ET can be calculated by above equation.

285
Comparing the data of three years in Table 2, DSR of the experiment site is relatively small, indicating that the soil moisture resource in the Pinus sylvestris var.mongolica forest is limited under this rain-fed conditions.In the dry years, the soil evaporation was relative large, but the measured total evapotranspiration decreased, indicating that the transpiration of vegetation was inhibited.Under this condition, Pinus sylvestris var.cannot grow well.In semiarid regions, precipitation varies considerably every year, and the year of 2016 may not be representative of the long-term 290 average behavior of DSR in this region as the precipitation of this year is higher than the average annual precipitation of 400 mm.Instead, it may be more representative of a wet year behavior, so as 2017 and 2018.

325
of Pinus sylvestris var.As extreme weather conditions happen more frequently worldwide (possibly due to the global warming effect), the arid region precipitation may change rapidly.Whether Pinus sylvestris var.mongolica can adapt to this trend is a question that still needs decades-long observational effort. 125

Figure 1 .
Figure 1.Geographic location of the experimental area and study site.

Figure 2 .
Figure 2. The schematic plot of a new lysimeter (on the right) with respect to the conventional lysimeter (on the left).

Figure 3 .
Figure 3. Annual precipitation and soil moisture of each layers in 2016.

Figure 4 .
Figure 4. Month soil moisture changes of every soil layers.

Figure 6 .
Figure 6.Soil moisture and DSR dynamic change in 2018.

2203. 1
Figure5A, it takes 2 days and 7 hours for the wetting point reaching the 60 cm soil layer, but for summer precipitation infiltration process, it takes only 1 day for wetting front to reach 60 cm soil layer.The February 26 th to March 27 th of 2016 snow melted soil moisture recharge process lasts for 29 days, and the soil moisture recharge depth reaches 160 cm.The soil moisture at 200 cm depth does not show any noticeable change, suggesting that DSR has not been generated.The start time of moisture recharge is set at when soil moisture content

Figure 7 .
Figure 7. Two recharge process (snow melt process and precipitation process) in 2016.

250a
designated depth, it leaves a crest signal.Based on the comparison between the number of crest signals and the number of precipitations, the minimum precipitation amount can be determined by the crest signals at different soil depths.Statistics of precipitation data from 2016 to 2018 and fluctuations in soil moisture content in each layer are shown in Table Hydrol.Earth Syst.Sci.Discuss., https://doi.org/10.5194/hess-2019-110Manuscript under review for journal Hydrol.Earth Syst.Sci. Discussion started: 18 March 2019 c Author(s) 2019.CC BY 4.0 License.

260Figure 3
Figure 3 shows that during the seasonal frozen-soil period, soil moisture is relatively stable.The monthly average values for soil moistures of different soil layers in January and December of 2016 are used as the start and end soil moisture values.Although the precipitation amount varies from 2016 to 2018, other environmental factors in this area are basically the same, and soil moistures are similar.To figure out the Pinus sylvestris var.mongolica water balance from 2016 to 2018, one has:

Table 1
Precipitation produced moisture increase signal and corresponding minimum precipitation intensity (data from 2016-2018).

Table 2
Water distribution in 2016 of the rain-feed Pinus sylvestris var.Table2, precipitation in the experimental area in 2016 is 466.4 mm and the DSR is 1.4 mm.Thanks to the heavy precipitation in the summer and the higher than multi-year average precipitation (multi-year average precipitation is 400 mm), all soil layer moisture content (δW) increases 38.056 mm within the upper 200 mm depth.The groundwater table is 8 m depth, beyond the root range of Pinus sylvestris var.mongolica.Therefore, Pinus sylvestris var.mongolica cannot absorb and utilize groundwater.Based on the real-time monitored precipitation, DSR 275 Hydrol.Earth Syst.Sci.Discuss., https://doi.org/10.5194/hess-2019-110Manuscriptunder review for journal Hydrol.Earth Syst.Sci.Discussion started: 18 March 2019 c Author(s) 2019.CC BY 4.0 License.andsoil moisture content change, the evapotranspiration of Pinus sylvestris var.mongolica can be calculated as 426.96 mm/year.2017 is a dry year with a precipitation of 309 mm, DSR of 0.4 mm, soil water storage decreased by 16 mm, evapotranspiration mount is 324.6 mm.The precipitation mount in 2018 is 472.2 mm.The DSR is 1.2 mm, the soil water storage capacity is increased by 54.747 mm, and the evapotranspiration is 416.253mm.Based on the redistribution data of precipitation in the shallow soil (200 cm depth) over the past three years, one can see that 280precipitation has a recharge effect on both shallow and deep soil layers.In the shallow soil layer, evapotranspiration in the dry year consumes stored water, but in the wet year precipitation water recharge the shallow soil layer.Deep soil layer infiltration has been recorded in the past three years with very small amount, indicating that under the existing vegetation cover and rain-fed conditions, the precipitation is barely able to support the shallow ecosystems, and only a small amount of precipitation can percolate into the deep soil layer.

4 Summary and Conclusions 295
As one of the four largest artificial shelter forest system, three North shelter forest system project which includes the Mu Us Sandy land investigated here, has a history of nearly 40 years of construction.On one hand, artificial shelter forest prevents the invasion of desert; on the other hand, construction of a vast artificial shelter forest may have detrimental effect on ecological environment in arid regions and may substantially change the evapotranspiration pattern in the region.Since precipitation is almost the only water sources for replenishing the groundwater system in 300 the Mu Us Sandy land, change of evapotranspiration will greatly affect the groundwater recharge in the region, which directly determines if the reforestation can be sustainable.This study uses a 30-years-old mature Pinus sylvestris var.mongolica forest of a 10-meter line spacing as the target of the experiment.The study uses a new lysimeter to monitor DSR and to accurately calculate water balance from 2016 to 2018.The following conclusions can be drawn from this study: rain-fed pine forest can meet the consumption of vegetations but the remaining amount of rain-fed infiltration that can percolate into deep soil layer is small.2.The experimental results show that the precipitation intensities are respectively 2.6 mm/d, 3.2 mm/d, 3.4 mm/d, 8.2 mm/d, 8.2 mm/d, and 13.2 mm/d when precipitation infiltrates into 20 cm, 40 cm, 80 cm, 120 cm, 160 cm, and 200 cm soil depths.Infiltration depth and precipitation intensity are not linearly related.315 3.In semi-arid areas, the annual precipitation varies greatly, the dry and wet years alternate, DSR mount is relatively small, soil water mount is limited.The growth of Pinus is affected by annual precipitation.Compared to the start of the year, soil moisture content increases 38.056 mm, -16 mm, 54.747 mm. 4. The precipitations in 2016-2018 are 466.4mm, 309 mm, 472.2 mm, and the associated DSR values are 1.4 mm, 0.2 mm, 1.2 mm, respectively.Under the current precipitation condition and reforestation design, the natural 320 recharging moisture of Pinus sylvestris var.mongolica can meet the growth needs, and have additional moisture for DSR which may eventually recharge groundwater.5. Calculation based on these dataset shows that the annual evaporation of Pinus sylvestris var.mongolica forest in Mu Us sandy land is 426.96mm, 324.6 mm, 416.253 mm for year 2016-2018, respectively.Pinus automatically adjusts its evapotranspiration in response to different precipitation amount, and this may affect the development To figure out the long-term behavior of DSR and SWC in the semiarid regions such as Mu Us Sandy land, one must carry out a multiyear (preferably a decade long) experiment.