我是Winbugs的新用户。我试图根据水营养素的测量来模拟土地利用系数。先前的分布（e0，e1，e2，e3，e4和k）是根据参考设置的。 Winbugs代码如下：
model
{
for (i in 1:S)
{
for (j in 1:W)
{
LL[i,j] ~ dnorm(mu[i,j], tau)
mu[i,j] <- e0[i,j]*Q[i,j] +
Q[i,j]*(e1[i,j]*Ac[i]+e2[i,j]*Acu[i]+e3[i,j]*Ad[i]+e4[i,j]*Ag[i])*exp(- k[i,j]*t0[i,j])+P[i,j]*exp(-k[i,j]*t1[i,j])
e0[i,j] ~ dnorm(e0.mu,e0.tau)
e1[i,j] ~ dnorm(e1.mu, e1.tau)
e2[i,j] ~ dnorm(e2.mu, e2.tau)
e3[i,j] ~ dnorm(e3.mu, e3.tau)
e4[i,j] ~ dnorm(e4.mu, e4.tau)
k[i,j] ~ dnorm(k.mu, k.tau)I(0,2)
}
}
# Choice of prior of random effects variances for reginal physiographic province
# Prior 1: Coast Plain (CP)
e0.mu ~ dnorm(0.34,7.69)
e1.mu ~ dnorm(0.56, 0.657)
e2.mu ~ dnorm(10.12, 0.465)
e3.mu ~ dnorm(0.02, 6.25)
e4.mu ~ dnorm(0.02,7.69)
e0.tau ~ dgamma(0.01,0.001)
e1.tau ~ dgamma(0.01,0.001)
e2.tau ~ dgamma(0.01,0.001)
e3.tau ~ dgamma(0.01,0.001)
e4.tau ~ dgamma(0.01,0.001)
k.mu ~ dnorm(0.5288, 19.8958)
k.tau ~ dgamma(0.01, 0.001)
tau ~ dgamma(0.01,0.01)
sigma <- 1/sqrt(tau)
}
# Initial values:
list(e0.tau = 0.34, e1.tau = 0.56, e2.tau = 10.12, e3.tau = 0.02, e4.tau = 0.02, k.tau=0.5288, tau =1)
# Data
list(S=1,W=68,
Q=structure(.Data=c(1679899.604,1140238.739,831143.8458,8408671.292,2799467.65,532017.5129,311427.4486,286463.204,499845.723,4431571.709,1352308.362,326476.7203,142180.1912,46101.46181,38311.57977,39099.346,18679.0187,102679.4153,100267.6604,78520.3203,35854.59685,4235.268648,1124.148612,1432.1626,6395.828609,22671.61925,49713.16967,49031.1077,38851.5676,62106.69949,146182.7689,151950.2647,216544.791,301893.5753,397369.4298,433260.784,507403.4991,785593.4744,582305.7251,1045098.195,2362058.204,2287228.068,671051.8377,694559.9749,1262452.995,687710.2186,1591227.233,1946579.929,6775923.725,3141655.847,937361.6256,1121568.543,561172.1346,667396.9998,342911.6248,186053.1583,166825.2309,122953.7924,103461.6847,45606.33054,22854.27754,7028.01481,2940.091499,65416.43176,6563.628011,1880.346074,39.63312539,6.211159109), .Dim=c(1,68)),
LL=structure(.Data=c(157910.5628,129987.2162,202799.0984,1202439.995,613083.4154,129280.2556,77856.86216,46979.96546,114464.6706,833135.4812,239358.58,59418.7631,20473.94753,599.3190036,0,508.291498,0,3696.458951,501.338302,0,0,0,0,0,0,0,546.8448664,0,0,558.9602954,146.1827689,607.8010587,649.6343729,905.6807258,794.7388596,18196.95293,4566.631492,366086.5591,127524.9538,251868.665,963719.7474,379679.8593,87907.79074,54175.67805,369898.7275,85276.06711,297559.4926,420461.2647,1748188.321,741430.7799,102172.4172,54956.85863,16273.9919,14682.734,2057.469749,5581.594749,9509.038163,6885.412375,4345.390757,45.60633054,0,0,0,719.5807494,0,0,0,0), .Dim=c(1,68)),
t0=structure(.Data=c(0.561147292,0.826731352,0.825358193,0.081581425,0.245043511,1.28941504,1.757443412,1.910598325,1.094970093,0.123503839,0.404727304,1.209585543,2.777472148,8.565921892,10.30762823,9.500551247,19.88676956,3.617719669,3.70473729,4.730817946,11.39338725,96.45322192,363.3908385,285.236681,60.72419134,17.13073583,7.812447341,7.921124741,13.5470244,8.474498543,3.600445787,3.463785571,2.43055089,3.302790585,2.509229909,2.301365124,1.965085499,1.829480331,2.468167059,1.375208393,0.60846418,0.73296968,2.498270225,2.413713554,1.327945541,2.475580275,1.069917493,0.874601565,0.251254577,0.541905904,1.00566429,0.840493539,1.679825237,1.412459322,2.00049031,3.687071957,4.112036163,5.579261681,5.290036785,12.00088039,23.94808223,77.87634668,186.1561512,6.036732827,60.16512824,210.0153405,9963.925811,59806.12215), .Dim=c(1,68)),
P=structure(.Data=c(70539.75634,70539.75634,150466.4078,150466.4078,150466.4078,150466.4078,133901.2987,133901.2987,133901.2987,133901.2987,133901.2987,69711.50088,69711.50088,69711.50088,69711.50088,131140.4472,131140.4472,131140.4472,131140.4472,131140.4472,75233.20392,75233.20392,75233.20392,75233.20392,141079.5127,141079.5127,141079.5127,141079.5127,70539.75634,70539.75634,70539.75634,70539.75634,70539.75634,70539.75634,70539.75634,70539.75634,70539.75634,137766.4908,137766.4908,137766.4908,137766.4908,137766.4908,137766.4908,137766.4908,137766.4908,206649.7363,206649.7363,206649.7363,206649.7363,206649.7363,68883.24542,68883.24542,68883.24542,68883.24542,62809.37208,62809.37208,62809.37208,62809.37208,69711.50088,69711.50088,69711.50088,69711.50088,69711.50088,66950.64936,66950.64936,66950.64936,66950.64936,66950.64936), .Dim=c(1,68)),
t1=structure(.Data=c(0.866639653,1.276809461,1.252839808,0.123835272,0.371960039,1.95724778,2.635565955,2.86524611,1.642081833,0.185213652,0.606952973,1.805176788,4.145079511,12.78372039,15.38303043,14.19962121,29.72296948,5.407080875,5.537138302,7.070728961,16.87872951,142.8905915,538.345229,422.5637801,89.39945574,25.22023638,11.50165239,11.66164958,20.50422317,12.82665508,5.449487779,5.242644455,3.678782616,5.108799683,3.881309647,3.55978168,3.039619956,2.772409383,3.74028045,2.084001991,0.922071571,1.082762133,3.690510631,3.565601288,1.961676158,3.71092251,1.603818284,1.311037524,0.376633421,0.812323009,1.553154695,1.298063875,2.594333395,2.181411678,3.036613578,5.596734318,6.241802216,8.468954683,7.933251653,17.99722914,35.91395876,116.7879697,279.1707609,9.009176768,89.79000581,313.4253877,14870.09141,89386.84278), .Dim=c(1,68)),
Ac=c(782.147065),
Acu=c(215.3012049),
Ad=c(2,009.40),
Ag=c(1,843.56))
在我用10000次迭代运行模型后，我可以得到每个系数的后验分布。然而，后验分布与先前的分布，甚至是数据范围有很大不同。我知道应该有一些差异，但一般来说范围应该是相同的。我非常感谢任何建议和意见。