【目的】利用易测林木因子模拟林下辐射方法体系中林下散射辐射测度指数尺度化加权散射荫蔽度与加权进界邻体散射荫蔽度的模拟准确性,以及这两种指数相对于开敞度的性能优劣,寻求基于简单、易测林木因子合理、准确估算林下任意点位辐射模型的方法。【方法】基于阔叶红松林(Pinus koraiensis)林分调查信息建立林分体素模型,模拟均质环境背景下林分内部样地与非均质环境背景下包含林缘样地的林下散射透射率,将其作为林下散射辐射测度指数模拟准确性的评价基准。分别以爬山算法对尺度化加权散射荫蔽度中的局域尺度圆半径、背景尺度圆半径及加权进界邻体散射荫蔽度中的局域临界值、背景临界值进行参数寻优。进一步评价3种散射辐射测度指数的性能优劣。【结果】均质环境背景下,尺度化加权散射荫蔽度在局域尺度圆半径为5.401 m、加权进界邻体散射荫蔽度在局域临界值为2.800时拟合林下散射透射率最优,此时开敞度、尺度化加权散射荫蔽度、加权进界邻体散射荫蔽度与林下散射透射率间的Pearson、Spearman相关系数分别为0.397与0.425、-0.716与-0.692、-0.730与-0.694,线性回归决定系数分别为0.158、0.514及0.533;相较于开敞度,2种林下散射辐射测度指数对林下散射透射率的方差解释率分别提升了237.3%与225.3%。非均质环境背景下,保持局域尺度圆半径及局域临界值与均质环境背景相同时,尺度化加权散射荫蔽度的背景尺度圆半径为15.521 m及加权进界邻体散射荫蔽度的背景临界值为0.875时拟合林下散射透射率最优,此时,尺度化加权散射荫蔽度、加权进界邻体散射荫蔽度与林下散射透射率间的Pearson、Spearman相关系数分别为-0.930与-0.719、-0.927与-0.820,线性回归决定系数分别为0.866与0.860。【结论】两种林下散射辐射成分测度指数在合理的参数选择下能较好地模拟林下散射辐射,并解释较小尺度内林木结构特征差异与较大尺度环境背景差异对林下散射辐射的影响,二者对林下散射辐射的模拟精度优于开敞度方法。在以林木因子模拟林下散射辐射时,需将大尺度环境背景的差异性纳入考量;同时,研究样点周围邻体林木分布的均匀性是模拟林下散射辐射时的重要影响因素。

【Objective】This study aims to evaluate the simulation accuracy of two measures of understory diffuse radiation: the scaled weighted diffuse shading degree and the weighted inside-boundary neighbor diffuse shading degree. These measures, based on easily measurable tree factors, are implemented within a forest understory radiation simulation system. The study also assesses these measures' performance relative to opening degree and seeks to identify model structural characteristics that can accurately estimate radiation at any understory location.【Method】A stand pixel model was created using survey data from a broad-leaved Korean pine forest (Pinus koraiensis). Understory diffuse transmittance was simulated for two scenarios: a homogeneous environmental background for research plots within the stand, and a heterogeneous environmental background for plots including the forest edge. The simulation accuracy of the diffuse radiation index measures was evaluated based on these scenarios. Optimization of the local-scale circle radius and background-scale circle radius for the scaled weighted diffuse shading degree, as well as the local critical value and background critical value for the weighted inside-boundary neighbor diffuse shading degree, were performed using a hill-climbing algorithm. The performance of the three indices-opening degree, scaled weighted diffuse shading degree, and weighted inside-boundary neighbor diffuse shading degree-were compared using Pearson and Spearman correlation coefficients and linear regression determination coefficients.【Result】Under a homogeneous environmental background, the scaled weighted diffuse shading degree showed the best fit for understory diffuse transmittance with a local-scale circle radius of 5.401 m. The weighted inside-boundary neighbor diffuse shading degree achieved the best fit with a local critical value of 2.800. Under these conditions, Pearson and Spearman correlation coefficients were 0.397 and 0.425 for opening degree, -0.716 and -0.692 for the scaled weighted diffuse shading degree, -0.730 and -0.694 for the weighted inside-boundary neighbor diffuse shading degree, respectively. The linear regression determination coefficients were 0.158 for opening degree, 0.514 for scaled weighted diffuse shading degree, and 0.533 for the weighted inside-boundary neighbor diffuse shading degree. Compared to the opening degree, the variance explanation rates for the weighted inside-boundary neighbor diffuse shading degree and the scaled weighted diffuse shading degree increased by 237.3% and 225.3%, respectively. In a heterogeneous environmental background, with the same local-scale circle radius and local critical value, the background-scale circle radius for the scaled weighted diffuse shading degree was 15.521 m, and the background critical value for the weighted inside-boundary neighbor diffuse shading degree was 0.875 when achieving the best fit for understory diffuse transmittance. In this case, the Pearson and Spearman correlation coefficients were -0.930 and -0.719 for the scaled weighted diffuse shading degree, -0.927 and -0.820 for the weighted inside-boundary neighbor diffuse shading degree. The linear regression determination coefficients were 0.866 and 0.860, respectively.【Conclusion】(1) With appropriate parameter selection, the scaled weighted diffuse shading degree and the weighted inside-boundary neighbor diffuse shading degree indices effectively simulate understory diffuse radiation and better account for variations in small-scale tree structure and large-scale environmental background compared to the opening degree method. (2) Accurate simulation of understory diffuse radiation requires consideration of large-scale environmental background differences, and the uniformity of neighboring tree distribution around research sites is an indispensable influencing factor in these simulations.