Neutral Model (QRule) Landscapes and Downscaling Code

The QRule_NeutralModels.zip dataset contains 37,620 binary rasters generated from the software QRule (https://www.umces.edu/qrule). Each raster (landscape) is a 512x512 raster. Each landscape was generated using one of 19 proportion values (P), where P ranges [5,95] in increments of 5. Each landscape was generated using one of 20 H values, where H ranges [0,95] in increments of 5. There are 380 unique P-H combinations, and we generated 99 repetitions of each for a total of 37,620 landscapes.<br>computeDownscaledMetrics.m is Matlab code for fitting scaling functions, downscaling those functions to predict values at 30 m resolution, and comparing predicted values to true values computed directly from the neutral model landscapes. The NeutralModels.xlsx contains a set of sample data for running the Matlab code.

QRule_NeutralModels.zip 数据集包含37,620张由QRule软件（https://www.umces.edu/qrule）生成的二值栅格（binary rasters）。每张栅格（景观）均为512×512分辨率。所有景观的生成均采用19个比例值P中的一个，P的取值范围为[5,95]，步长为5；同时采用20个H值中的一个，H的取值范围为[0,95]，步长为5。总计存在380种独特的P-H组合，每种组合生成99次重复，最终共得到37,620张景观栅格。<br>computeDownscaledMetrics.m 为Matlab代码，可用于拟合缩放函数、将这些函数降尺度至30米分辨率以完成预测，并将预测值与直接从中性模型景观中计算得到的真实值进行比对。NeutralModels.xlsx 包含了用于运行该Matlab代码的示例数据集。