Subsurface temperature at Lomonosovfonna, Svalbard, April 2012-2016

The dataset contains subsurface temperature measurements done at Lomonosovfonna, Svalbard, during April 2012 - 2016.All measurements are done at the site with coordinates: 78.8235 N, 17.432 E.The data is contained in four cells of a matlab structure containing data from installations deployed in April 2012 - cell 1, April 2013 - cell 2, April 2014 - cell 3 and April 2015 - cell 4. In 2012-2014 nine thermistor strings were installed in each year. The nine T-strings were arranged in a 3*3 square grid with a 3 m spacing between neighboring strings. In 2015 one t-string was installed.Hardware: Campbell Scientific CR10X data loggers in combination with several relay multiplexers (AM416 of AM16/32B) were used for temperature measurements. For that a reference temperature stable resistor (Rr Ohm) was connected is series with thermistors. Known excitation voltage (Ue) was supplied to the circuit and the voltage was measured (Um) at the leads of the reference resistor.The resistance of a thermistor (Rt) was then calculated as:Rt = Ue * Rr / Um - Rr.The resistance was then converted to temperature values provided by the manufacturer of thermistors.Technical information is contained in the variables: LF{N}.T.system.The raw temperature measurements along with the time stamps and depths are contained in the variables LF{N}.T.system.T_raw, LF{N}.T.system.t_raw and LF{N}.T.system.z_raw.After unpacking the data was subjected to the following post-processing steps:- delete data from sensors that were left above the snow surface- for the sensors installed in April 2013: delete data after 2013 July 12- reset temperature values outsides of the range [-40 +10] degC to NaN- for the sensors installed in April 2015: correct values from one of the sensors by linear interpolation in time between the following time points: 2015 November 15 02:00 and 2015 November 15 14:00, 2015 December 18 15:00 and 2015 December 19 21:00- introduce corrections to depths of sensors to match temperature distributions measured at different T-strings during the periods dominated by conductive heat exchange in the firn packcorrections are contained in the variable LF{N}.T.system.z_off and are given in meters.- delete data from sensors that are deemed erroneous.For the sensors installed in April 2012 that is: sensor 1 in T-string 9.For the sensors installed in April 2013 that is: sensors 1 and 2 in T-string 2, sensors 1-6 in T-string 3, sensors 1-6 in T-string 4, sensors 1-5 in T-string 5, sensors 1-7 in T-string 6.For the sensors installed in April 2014 that is: sensors 1 in T-string 1, sensor 1 in T-string 7, sensor 1 in T-string 9.- apply offsets for individual sensors defined as the mode during the time period, when the temperature is expected to be at 0 degC. For the sensors installed in April 2012 and 2015 that is the entire measurement period. For the sensors installed in April 2014 the periods are defined based on subjective data analysis and are different for individual sensors. For the sensors installed in April 2013 and some sensors installed in April 2014 the offsets are set to 0 degC. The applied temperature offsets are contained in the variables: LF{N}.T.system.off. The relation between the number of temperature values equal to the offset and the total number of values during the calibration time is saved in the variable LF{N}.T.system.f.After the above described post-processing steps the data was saved in the variable LF{N}.T.T (temperature values), LF{N}.T.z (depths of sensors) and LF{N}.T.t (time stamps).Data interpolated on a regular grid is contained in the variables: LF{N}.T.T_i (temperature values) and LF{N}.T.z_i (depth vectors).Data laterally averaged across all T-strings is contained in the variables: LF{N}.T.T_a (temperature values) and LF{N}.T.z_a (depth vectors).The standard deviation in interpolated temperature values belonging to the same depth but coming from different T-strings are contained in the variables LF{N}.T.T_sd.The data was used in the following publications:1) Marchenko, S., Cheng, G., Lötstedt, P., Pohjola, V., Pettersson, R., van Pelt, W., Reijmer, C., (2019). Thermal conductivity of firn at Lomonosovfonna, Svalbard, derived from subsurface temperature measurements, The Cryosphere Discussions, doi: 10.5194/tc-2018-294;2) Marchenko, S., van Pelt, W., Claremar, B., Pohjola, V., Pettersson, R., Machguth, H., Reijmer, C., (2017). Parameterizing Deep Water Percolation Improves Subsurface Temperature Simulations by a Multilayer Firn Model, Frontiers in Earth Science, doi: 10.3389/feart.2017.00016;3) Marchenko, S., Pohjola, V., Pettersson, R., van Pelt, W., Vega, C., Machguth, H., Bøggild C., Isaksson, E., (2017). A plot-scale study of firn stratigraphy at Lomonosovfonna, Svalbard, using ice cores, borehole video and GPR surveys in 2012-14, Journal of Glaciology, doi: 10.1017/jog.2016.118;本数据集包含2012年4月至2016年4月期间，在斯瓦尔巴群岛（Svalbard）洛蒙诺索夫冰穹（Lomonosovfonna）获取的地下温度测量数据。所有测量均在坐标为78.8235°N、17.432°E的测点完成。 数据存储于一个Matlab结构体（Matlab structure）的四个单元中，各单元对应不同的设备部署时间：单元1对应2012年4月部署的监测设备，单元2对应2013年4月，单元3对应2014年4月，单元4对应2015年4月。2012至2014年每年均部署9条热敏电阻串（thermistor string，下文简称T-string），该9条测串按3×3方格网格排布，相邻测串间距为3米；2015年仅部署1条热敏电阻串。 测量硬件采用Campbell Scientific CR10X数据记录仪（data logger），搭配多款中继多路复用器（relay multiplexer，型号为AM416或AM16/32B）开展温度测量。测量时将参考温度稳定电阻器（reference temperature stable resistor，记为Rr欧姆）与热敏电阻串联，向电路输入已知激励电压（Ue），并测量参考电阻两端的电压（Um）。据此可通过如下公式计算热敏电阻阻值Rt： Rt = Ue × Rr / Um - Rr 随后将热敏电阻阻值转换为温度值，转换方法由热敏电阻生产商提供。 技术相关信息存储于变量LF{N}.T.system中。原始温度测量数据、对应时间戳与传感器埋深分别存储于变量LF{N}.T.system.T_raw、LF{N}.T.system.t_raw与LF{N}.T.system.z_raw中。 数据解压后需经过以下后处理步骤： 1. 删除位于雪面以上的传感器采集数据； 2. 针对2013年4月部署的传感器，删除2013年7月12日之后的全部采集数据； 3. 将温度值超出[-40, 10]℃范围的测点标记为非数值（NaN）并重置； 4. 针对2015年4月部署的传感器，对其中1条传感器的测量值进行时间线性插值修正，修正区间为2015年11月15日02:00至14:00，以及2015年12月18日15:00至2015年12月19日21:00； 5. 对传感器埋深引入修正量，以匹配粒雪层（firn pack）中以传导换热为主的时段内，不同T-string测得的温度分布。修正量存储于变量LF{N}.T.system.z_off中，单位为米； 6. 删除被判定为异常的传感器采集数据： - 2012年4月部署的传感器：T-string 9中的传感器1； - 2013年4月部署的传感器：T-string 2中的传感器1、2，T-string 3中的传感器1-6，T-string 4中的传感器1-6，T-string 5中的传感器1-5，T-string 6中的传感器1-7； - 2014年4月部署的传感器：T-string 1中的传感器1，T-string 7中的传感器1，T-string 9中的传感器1； 7. 为各传感器应用温度偏移量：偏移量定义为温度理论值应为0℃的时段内的众数。其中，2012年与2015年部署的传感器的校准时段为整个测量周期；2014年部署的传感器的校准时段需通过主观数据分析确定，且各传感器的校准时段各不相同；2013年部署的传感器与部分2014年部署的传感器的偏移量设为0℃。应用后的温度偏移量存储于变量LF{N}.T.system.off中。校准时段内温度值等于偏移量的样本数占总样本数的比例，存储于变量LF{N}.T.system.f中。 经过上述后处理流程后，数据分别存储于以下变量：LF{N}.T.T（温度值）、LF{N}.T.z（传感器埋深）与LF{N}.T.t（时间戳）。 按规则网格插值得到的数据存储于变量LF{N}.T.T_i（温度值）与LF{N}.T.z_i（深度向量）。 所有T-string的横向平均数据存储于变量LF{N}.T.T_a（温度值）与LF{N}.T.z_a（深度向量）。 同一埋深下不同T-string的插值温度值的标准差，存储于变量LF{N}.T.T_sd中。 本数据集已用于以下学术出版物： 1. Marchenko, S., Cheng, G., Lötstedt, P., Pohjola, V., Pettersson, R., van Pelt, W., Reijmer, C., (2019). 基于地下温度测量数据推导的斯瓦尔巴群岛洛蒙诺索夫冰穹粒雪热导率，《冰冻圈讨论（The Cryosphere Discussions）》，doi: 10.5194/tc-2018-294； 2. Marchenko, S., van Pelt, W., Claremar, B., Pohjola, V., Pettersson, R., Machguth, H., Reijmer, C., (2017). 深层水渗流参数化改进多层粒雪模型的地下温度模拟效果，《前沿地球科学（Frontiers in Earth Science）》，doi: 10.3389/feart.2017.00016； 3. Marchenko, S., Pohjola, V., Pettersson, R., van Pelt, W., Vega, C., Machguth, H., Bøggild C., Isaksson, E., (2017). 2012-2014年利用冰芯、钻孔视频与探地雷达调查斯瓦尔巴群岛洛蒙诺索夫冰穹的粒雪地层学——地块尺度研究，《冰川学期刊（Journal of Glaciology）》，doi: 10.1017/jog.2016.118；