Altimetric observation of wave attenuation through the Antarctic marginal ice zone using ICESat-2

Data from all ICESat-2 tracks underlying the publication Brouwer et al., 2022:
Brouwer, J., Fraser, A. D., Murphy, D. J., Wongpan, P., Alberello, A., Kohout, A., Horvat, C., Wotherspoon, S., Massom, R. A., Cartwright, J., and Williams, G. D.: Altimetric observation of wave attenuation through the Antarctic marginal ice zone using ICESat-2, The Cryosphere, 16, 2325–2353, https://doi.org/10.5194/tc-16-2325-2022, 2022.

ICESat-2 is a high-precision laser altimeter in a polar orbit. Due to its narrow laser footprint (~17 m) and high measurement repetition rate (firing every 70 cm) it has the capability to resolve the passage of waves (of wavelength ~100 to 1000 m) through sea ice in the Antarctic marginal ice zone. This dataset provides the processed data underlying the ICESat-2 tracks published in Brouwer et al., 2022. The processing is described in that publication, but involves manual and automated filtering to remove inappropriate (cloud-affected) tracks, then a series of steps to convert segment heights into estimates of marginal ice zone width.

Rout
A brief description of the dataset fields is provided below. For more information please see methods section of paper, and the following code: https://github.com/Jill-Brouwer/Brouwer-etal-2022-MIZ-code/blob/main/R-gam3batch.R
• Rows describe FFT and FIRF techniques for calculating significant wave height
o hm0bc: FFT with boxcar window
o hm0hann: FFT with hanning window
o std: 4 times the standard deviation of height
o FXXXX refer to the FIRFs and the wavelengths ranging from 1500m to 38m
• Columns describe the parameters from the segmented regression fits and resulting MIZ width estimate.
o DC_slope_lin: Slope derived from linear regression fit to relevant significant wave height estimate (row value, m units) and corrected distance of the section midpoint from ice edge (km units) for the outer (attenuation-dominated) segment.
o DC_slope_lin_error: standard error in DC_slope_lin generated by the regression function
o DC_int_lin: Y intercept derived from segmented linear regression fit to relevant significant wave height estimate (row value, m units) and corrected distance of the section midpoint from ice edge (km units) for the outer (attenuation-dominated) segment.
o DC_int_lin_err: standard error in DC_int_lin generated by the regression function
o DC_bp_lin: break point estimated from R segmented function in corrected distance from ice edge (km units) and linear significant wave height estimates (m)
o DC_miz_lin: Miz width estimated using the attenuation parameters DC_slope_lin and DC_int_lin
o DC_RMSE_lin: RMSE of linear regression fit to relevant significant wave height estimate (row value, m units) and corrected distance of the section midpoint from ice edge (km units) for the outer (attenuation-dominated) segment.
o DC_miz_lin_err: error in MIZ width estimate calculated by propagating errors in the DC_slope_lin_err and DC_int_lin_err
o DC_slope_exp: Slope derived from exponential regression fit to relevant significant wave height estimate (row value, m units) and corrected distance of the section midpoint from ice edge (km units) for the outer (attenuation-dominated) segment.
o DC_slope_exp_err: standard error in DC_slope_exp generated by the regression function
o DC_exp_lin: Y intercept derived from exponential segmented regression fit to relevant significant wave height estimate (row value) and corrected distance of the section midpoint from ice edge (km units) for the outer/attenuation dominated segment.
o DC_int_exp_err: standard error in DC_exp_lin generated by the regression function
o DC_bp_exp: break point estimated from R segmented function in corrected distance from ice edge (km units) and log-transformed significant wave height estimates; exponential fit.
o DC_miz_exp: Miz width estimated using the attenuation parameters DC_slope_exp and DC_exp_lin
o DC_RMSE_exp: RMSE of exponential regression fit to relevant significant wave height estimate (row value) and corrected distance of the section midpoint from ice edge (km units) for the outer (attenuation-dominated) segment.
o DC_miz_exp_err: error in MIZ width estimate calculated by propagating errors DC_slope_exp_err and DC_int_exp_err.
o D_slope_lin: Slope derived from linear regression fit to relevant significant wave height estimate (row value, m units) and distance of the section midpoint from ice edge (km units) for the outer (attenuation-dominated) segment.
o D_slope_lin_err: standard error in D_slope_lin generated by the regression function
o D_int_lin: Y intercept derived from segmented linear regression fit to relevant significant wave height estimate (row value, m units) and distance of the section midpoint from ice edge (km units) for the outer (attenuation-dominated) segment.
o D_int_lin_err: standard error in D_int_lin generated by the regression function
o D_bp_lin: break point estimated from R segmented function in distance from ice edge (km units) and significant wave height estimates (m)
o D_miz_lin: Miz width estimated using the attenuation parameters D_slope_lin and D_int_lin
o D_RMSE_lin: RMSE of linear regression fit to relevant significant wave height estimate (row value, m units) and distance of the section midpoint from ice edge (km units) for the outer/attenuation dominated segment.
o D_miz_lin_err: error in MIZ width estimate calculated by propagating errors in the D_slope_lin_err and D_int_lin_err
o D_slope_exp: Slope derived from exponential regression fit to relevant significant wave height estimate (row value, m units) and distance of the section midpoint from ice edge (km units) for the outer/attenuation dominated segment.
o D_slope_exp_err: standard error in D_slope_exp generated by the regression function
o D_exp_lin: Y intercept derived from exponential segmented regression fit to relevant significant wave height estimate (row value) and corrected distance of the section midpoint from ice edge (km units) for the outer (attenuation-dominated) segment.
o D_int_exp_err: standard error in DC_exp_lin generated by the regression function
o D_bp_exp: break point estimated from R segmented function in corrected distance from ice edge (km units) and log-transformed significant wave height estimates; exponential fit.
o D_miz_exp: Miz width estimated using the attenuation parameters D_slope_exp and D_exp_lin
o D_RMSE_exp: RMSE of exponential regression fit to relevant significant wave height estimate (row value) and corrected distance of the section midpoint from ice edge (km units) for the outer (attenuation-dominated) segment.
o D_miz_exp_err: error in MIZ width estimate calculated by propagating errors D_slope_exp_err and D_int_exp_err.
• The final column parameters provide some extra values from the broken regression function that describe the inner ice-structure dominated fit:
o DC_sgslop_lin: Slope derived from linear regression fit to relevant significant wave height estimate (row value, m units) and corrected distance of the section midpoint from ice edge (km units) for the inner (ice-structure-dominated) segment.
o DC_segslop_lin_err: standard error in DC_sgslope_lin generated by the regression function
o DC_segint_lin: Y intercept derived from segmented linear regression fit to relevant significant wave height estimate (row value, m units) and corrected distance of the section midpoint from ice edge (km units) for the inner (ice-structure-dominated) segment.
o DC_bp_lin_err: standard error in DC_bp_lin generated by the regression function
o DC_segslop_exp: Slope derived from exponential regression fit to relevant significant wave height estimate (row value, m units) and corrected distance of the section midpoint from ice edge (km units) for the inner (ice-structure-dominated) segment.
o DC_segslop_exp_err: standard error in DC_segslope_exp generated by the regression function
o DC_segint_exp: Y intercept derived from segmented regression fit to log-transformed significant wave height estimate (row value, m units) and corrected distance of the section midpoint from ice edge (km units) for the inner (ice-structure-dominated) segment.
o DC_bp_exp_err: standard error in DC_bp_exp generated by the regression function
o D_segslop_lin: Slope derived from linear regression fit to relevant significant wave height estimate (row value, m units) and distance of the section midpoint from ice edge (km units) for the inner (ice-structure-dominated) segment.
o D_segslop_lin_err: standard error in D_segslope_lin generated by the regression function
o D_segint_lin: Y intercept derived from segmented linear regression fit to relevant significant wave height estimate (row value, m units) and distance of the section midpoint from ice edge (km units) for the inner (ice-structure-dominated) segment.
o D_bp_lin_err: standard error in D_bp_lin generated by the regression function
o D_segslop_exp: Slope derived from exponential regression fit to relevant significant wave height estimate (row value, m units) and distance of the section midpoint from ice edge (km units) for the inner (ice-structure-dominated) segment.
o D_segslop_exp_err: standard error in D_segslope_exp generated by the regression function
o D_segint_exp: Y intercept derived from segmented regression fit to log-transformed significant wave height estimate (row value, m units) and distance of the section midpoint from ice edge (km units) for the inner (ice-structure-dominated) segment.
o D_bp_exp_err: standard error in D_bp_exp generated by the regression function
• The final column, SIC_dist, lists the MIZ width derived from sea ice concentration data in km.

Rawheight
Describes distance from ice edge in meters at 8m intervals and the matching height values interpolated from IS-2 altimeter heights with cubic spline.

Csv (mizfind)
This is the data that was used in fitting the linear and exponential fits to significant wave height and distance from the ice edge for each 6.25km section midpoint, for final estimation of MIZ width. For more information please see methods section of paper and source code available here: https://github.com/Jill-Brouwer/Brouwer-etal-2022-MIZ-code/tree/main.

The following variables are for section midpoints with suitable significant wave height estimates:
• Dist: midpoint distance from ice edge (in kilometers)
• CDist: midpoint corrected distance from ice edge (in kilometers)
• Hm0bc: significant wave height from spectral analysis of Icesat-2 interpolated height data using boxcar window
• Hm0hann: significant wave height from spectral analysis of Icesat-2 interpolated height data using hann window
• Std: significant wave height from the standard deviation (SD) of the Icesat-2 interpolated height data
• FXXXX: significant wave height from spatial domain filters where XXXX corresponds to the wavelength used (in meters).
• hm0bcerr: hm0bc error
• hm0hannerr: hm0hann error
• stderr: std error
• FXXXXerr: error in FXXXX
• Mplat, mplon: Latitude and longitude of section midpoint
• Mpsic: sea ice concentration of section midpoint
• Good_mpinds: lookup table of corresponding section midpoint indexes that had acceptable significant wave height estimates

The below variables are for all section midpoints from the ice edge - including those that do not have suitable significant wave height estimates.
• all_sic_dist: Distance from ice edge in km for section midpoint
• all_sic_corrdist: Corrected distance from ice edge in km for section midpoint
• all_sic_vals: sea ice concentration value for section midpoint
• all_lat, all_lon: latitude and longitude of all section midpoints
• beam2mp_time: midpoint time extracted from IS-2 dataset

本数据集涵盖了发表于Brouwer等人2022年论文中的全部ICESat-2轨迹数据：
Brouwer, J.、Fraser, A. D.、Murphy, D. J.、Wongpan, P.、Alberello, A.、Kohout, A.、Horvat, C.、Wotherspoon, S.、Massom, R. A.、Cartwright, J. 与 Williams, G. D.：《利用ICESat-2开展南极边际冰区（Marginal Ice Zone, MIZ）海浪衰减的测高观测》，《冰冻圈》，16卷，2325–2353页，https://doi.org/10.5194/tc-16-2325-2022，2022年。

ICESat-2是一台部署于极轨轨道的高精度激光测高仪。凭借其狭窄的激光足迹（约17米）与极高的测量重复频率（每70厘米发射一次激光），该仪器能够解析波长约100至1000米的海浪穿过南极边际冰区海冰的过程。本数据集提供了Brouwer等人2022年论文中发表的ICESat-2轨迹的经处理数据。处理流程已在该论文中详述，包括手动与自动滤波以剔除受云影响的不合格轨迹，以及一系列将分段高度转换为边际冰区宽度估算值的步骤。

### Rout
下文将对该数据集的字段进行简要说明。如需获取更多信息，请参阅论文的方法章节以及以下代码：https://github.com/Jill-Brouwer/Brouwer-etal-2022-MIZ-code/blob/main/R-gam3batch.R
• 行数据对应用于计算有效波高的快速傅里叶变换（Fast Fourier Transform, FFT）与有限长冲激响应滤波（Finite Impulse Response Filter, FIRF）方法：
o hm0bc：采用矩形窗（boxcar window）的FFT计算结果
o hm0hann：采用汉宁窗（hanning window）的FFT计算结果
o std：高度标准差的4倍
o FXXXX：对应FIRF滤波方法，其中XXXX代表所用波长（范围为1500米至38米）
• 列数据对应分段回归拟合得到的参数以及最终的边际冰区宽度估算值：
o DC_slope_lin：针对外部（衰减主导）分段，将有效波高估算值（行数据，单位：米）与剖面中点到冰缘的校正距离（单位：千米）进行线性回归拟合得到的斜率
o DC_slope_lin_error：回归函数计算得到的DC_slope_lin的标准误
o DC_int_lin：针对外部（衰减主导）分段，将有效波高估算值（行数据，单位：米）与剖面中点到冰缘的校正距离（单位：千米）进行分段线性回归拟合得到的Y轴截距
o DC_int_lin_err：回归函数计算得到的DC_int_lin的标准误
o DC_bp_lin：通过R分段回归函数估算得到的拐点，对应校正后的冰缘距离（单位：千米）与线性有效波高估算值（单位：米）
o DC_miz_lin：利用衰减参数DC_slope_lin与DC_int_lin估算得到的边际冰区宽度
o DC_RMSE_lin：针对外部（衰减主导）分段，将有效波高估算值（行数据，单位：米）与剖面中点到冰缘的校正距离（单位：千米）进行线性回归拟合的均方根误差（Root Mean Square Error, RMSE）
o DC_miz_lin_err：通过传递DC_slope_lin_error与DC_int_lin_err的误差计算得到的边际冰区宽度估算误差
o DC_slope_exp：针对外部（衰减主导）分段，将有效波高估算值（行数据，单位：米）与剖面中点到冰缘的校正距离（单位：千米）进行指数回归拟合得到的斜率
o DC_slope_exp_err：回归函数计算得到的DC_slope_exp的标准误
o DC_exp_lin：针对外部/衰减主导分段，将有效波高估算值（行数据）与剖面中点到冰缘的校正距离（单位：千米）进行指数分段回归拟合得到的Y轴截距
o DC_int_exp_err：回归函数计算得到的DC_exp_lin的标准误
o DC_bp_exp：通过R分段回归函数估算得到的拐点，对应校正后的冰缘距离（单位：千米）与对数转换后的有效波高估算值，为指数拟合结果
o DC_miz_exp：利用衰减参数DC_slope_exp与DC_exp_lin估算得到的边际冰区宽度
o DC_RMSE_exp：针对外部（衰减主导）分段，将有效波高估算值（行数据）与剖面中点到冰缘的校正距离（单位：千米）进行指数回归拟合的均方根误差
o DC_miz_exp_err：通过传递DC_slope_exp_err与DC_int_exp_err的误差计算得到的边际冰区宽度估算误差
o D_slope_lin：针对外部（衰减主导）分段，将有效波高估算值（行数据，单位：米）与剖面中点到冰缘的距离（单位：千米）进行线性回归拟合得到的斜率
o D_slope_lin_err：回归函数计算得到的D_slope_lin的标准误
o D_int_lin：针对外部（衰减主导）分段，将有效波高估算值（行数据，单位：米）与剖面中点到冰缘的距离（单位：千米）进行分段线性回归拟合得到的Y轴截距
o D_int_lin_err：回归函数计算得到的D_int_lin的标准误
o D_bp_lin：通过R分段回归函数估算得到的拐点，对应冰缘距离（单位：千米）与有效波高估算值（单位：米）
o D_miz_lin：利用衰减参数D_slope_lin与D_int_lin估算得到的边际冰区宽度
o D_RMSE_lin：针对外部/衰减主导分段，将有效波高估算值（行数据，单位：米）与剖面中点到冰缘的距离（单位：千米）进行线性回归拟合的均方根误差
o D_miz_lin_err：通过传递D_slope_lin_err与D_int_lin_err的误差计算得到的边际冰区宽度估算误差
o D_slope_exp：针对外部/衰减主导分段，将有效波高估算值（行数据，单位：米）与剖面中点到冰缘的距离（单位：千米）进行指数回归拟合得到的斜率
o D_slope_exp_err：回归函数计算得到的D_slope_exp的标准误
o D_exp_lin：针对外部（衰减主导）分段，将有效波高估算值（行数据）与剖面中点到冰缘的校正距离（单位：千米）进行指数分段回归拟合得到的Y轴截距
o D_int_exp_err：回归函数计算得到的DC_exp_lin的标准误（原文此处存在笔误，应为D_exp_lin）
o D_bp_exp：通过R分段回归函数估算得到的拐点，对应校正后的冰缘距离（单位：千米）与对数转换后的有效波高估算值，为指数拟合结果
o D_miz_exp：利用衰减参数D_slope_exp与D_exp_lin估算得到的边际冰区宽度
o D_RMSE_exp：针对外部（衰减主导）分段，将有效波高估算值（行数据）与剖面中点到冰缘的校正距离（单位：千米）进行指数回归拟合的均方根误差
o D_miz_exp_err：通过传递D_slope_exp_err与D_int_exp_err的误差计算得到的边际冰区宽度估算误差
• 最后几列参数来自分段回归函数，用于描述内部（海冰结构主导）分段的拟合结果：
o DC_sgslop_lin：针对内部（海冰结构主导）分段，将有效波高估算值（行数据，单位：米）与剖面中点到冰缘的校正距离（单位：千米）进行线性回归拟合得到的斜率
o DC_segslop_lin_err：回归函数计算得到的DC_sgslop_lin的标准误
o DC_segint_lin：针对内部（海冰结构主导）分段，将有效波高估算值（行数据，单位：米）与剖面中点到冰缘的校正距离（单位：千米）进行分段线性回归拟合得到的Y轴截距
o DC_bp_lin_err：回归函数计算得到的DC_bp_lin的标准误
o DC_segslop_exp：针对内部（海冰结构主导）分段，将有效波高估算值（行数据，单位：米）与剖面中点到冰缘的校正距离（单位：千米）进行指数回归拟合得到的斜率
o DC_segslop_exp_err：回归函数计算得到的DC_segslop_exp的标准误
o DC_segint_exp：针对内部（海冰结构主导）分段，将对数转换后的有效波高估算值（行数据，单位：米）与剖面中点到冰缘的校正距离（单位：千米）进行分段回归拟合得到的Y轴截距
o DC_bp_exp_err：回归函数计算得到的DC_bp_exp的标准误
o D_segslop_lin：针对内部（海冰结构主导）分段，将有效波高估算值（行数据，单位：米）与剖面中点到冰缘的距离（单位：千米）进行线性回归拟合得到的斜率
o D_segslop_lin_err：回归函数计算得到的D_segslop_lin的标准误
o D_segint_lin：针对内部（海冰结构主导）分段，将有效波高估算值（行数据，单位：米）与剖面中点到冰缘的距离（单位：千米）进行分段线性回归拟合得到的Y轴截距
o D_bp_lin_err：回归函数计算得到的D_bp_lin的标准误
o D_segslop_exp：针对内部（海冰结构主导）分段，将有效波高估算值（行数据，单位：米）与剖面中点到冰缘的距离（单位：千米）进行指数回归拟合得到的斜率
o D_segslop_exp_err：回归函数计算得到的D_segslop_exp的标准误
o D_segint_exp：针对内部（海冰结构主导）分段，将对数转换后的有效波高估算值（行数据，单位：米）与剖面中点到冰缘的距离（单位：千米）进行分段回归拟合得到的Y轴截距
o D_bp_exp_err：回归函数计算得到的D_bp_exp的标准误
• 最后一列SIC_dist为通过海冰密集度数据计算得到的边际冰区宽度，单位为千米。

### Rawheight
该数据集记录了以8米为间隔的冰缘距离数据，以及通过三次样条插值从ICESat-2测高数据中得到的匹配高度值。

### Csv (mizfind)
该数据集用于针对每个6.25千米剖面中点的有效波高与冰缘距离开展线性与指数拟合，最终实现边际冰区宽度的估算。如需获取更多信息，请参阅论文的方法章节以及以下开源代码：https://github.com/Jill-Brouwer/Brouwer-etal-2022-MIZ-code/tree/main.

以下变量对应具备有效有效波高估算值的剖面中点：
• Dist：剖面中点到冰缘的距离（单位：千米）
• CDist：剖面中点到冰缘的校正距离（单位：千米）
• Hm0bc：通过矩形窗对ICESat-2插值高度数据进行频谱分析得到的有效波高
• Hm0hann：通过汉宁窗对ICESat-2插值高度数据进行频谱分析得到的有效波高
• Std：基于ICESat-2插值高度数据的标准差计算得到的有效波高
• FXXXX：通过空间域滤波得到的有效波高，其中XXXX代表所用波长（单位：米）
• hm0bcerr：hm0bc的估算误差
• hm0hannerr：hm0hann的估算误差
• stderr：std的标准误
• FXXXXerr：FXXXX的估算误差
• Mplat, mplon：剖面中点的纬度与经度
• Mpsic：剖面中点的海冰密集度
• Good_mpinds：具备合格有效波高估算值的对应剖面中点索引查找表

以下变量对应来自冰缘的全部剖面中点，包括不具备有效有效波高估算值的中点：
• all_sic_dist：剖面中点到冰缘的距离（单位：千米）
• all_sic_corrdist：剖面中点到冰缘的校正距离（单位：千米）
• all_sic_vals：剖面中点的海冰密集度数值
• all_lat, all_lon：全部剖面中点的纬度与经度
• beam2mp_time：从ICESat-2数据集中提取得到的中点时间